University of Florida Levin College of Law
UF Law Scholarship Repository
Faculty Publications
Faculty Scholarship
Summer 1994
Double Jeopardy, the Federal Sentencing
Guidelines, and the Subsequent-Prosecution
Dilemma
Elizabeth T. Lear
University of Florida Levin College of Law, [email protected]
Follow this and additional works at: http://scholarship.law.ufl.edu/facultypub
Part of the Constitutional Law Commons, Criminal Law Commons, and the Criminal Procedure
Commons
Recommended Citation
Elizabeth T. Lear, Double Jeopardy, the Federal Sentencing Guidelines, and the Subsequent-Prosecution Dilemma, 60 Brook. L. Rev. 725
(1994), available at http://scholarship.law.ufl.edu/facultypub/449
This Article is brought to you for free and open access by the Faculty Scholarship at UF Law Scholarship Repository. It has been accepted for inclusion
in Faculty Publications by an authorized administrator of UF Law Scholarship Repository. For more information, please contact [email protected].
DOUBLE JEOPARDY, THE FEDERAL
SENTENCING GUIDELINES, AND THE
SUBSEQUENT-PROSECUTION DILEMMA
Elizabeth T. Lear*
INTRODUCTION
The choice to embrace a real-offense regime probably constitutes the single most controversial decision made by the
Federal Sentencing Commission in drafting the Federal Sentencing Guidelines ("Guidelines").1 Real-offense sentencing
* Associate Professor of Law, University of Florida; Visiting Professor of Law,
California Western School of Law. J.D., 1987, University of Michigan; B.A., 1981,
University of North Carolina at Chapel Hill. I am deeply indebted to Lyrissa
Barnett, Charles Collier, Kenwyn Fuller, William Hazzard, Jerold Israel, Andrew
Kind, Marc Miller, Michael Seigel, Christopher Slobogin and David Yellen for their
insightful comments, and to my excellent research assistants, Bryan Aylstock,
Deborah Ben-David, Michelle Harrington and Michael Winter for their fine work
and wonderful sense of loyalty. I also wish to thank California Western School of
Law for its generous research support for this project.
' In 1984, Congress passed the Sentencing Reform Act, which created the
United States Sentencing Commission (the "Commission") to design sentencing
guidelines for the federal courts ("Guidelines"). See Sentencing Reform Act of 1984,
18 U.S.C. §§ 3551-3586; 28 U.S.C. §§ 991-998. The resulting guidelines became law
on November 1, 1987. See UNITED STATES SENTENCING COMMISSION, GUIDELINES
MANUAL 1 (1992) [hereinafter U.S.S.G.].
The Guidelines rely on a real-offense as opposed to a charge-offense model.
"Real-offense" sentencing refers to a system that is theoretically unconstrained by
the offense of conviction or indictment and that bases the sentence on what "actually" happened as determined by the sentencing court. In the pre-Guidelines era,
the real-offense inquiry was quite broad, encompassing a variety of factors, including the circumstances of the offense of conviction, nonconviction offenses committed
before, during or after the conviction offense, and a host of biographic components
such as good works and employment history. See Kevin R. Reitz, Sentencing Facts:
Travesties of Real-Offense Sentencing, 45 STAN. L. REV. 523, 526-28 (1993). And although the Guidelines technically adopt a modified real-offense system by excluding much unrelated criminal activity and other information from the sentencing inquiry, a vast range of conduct remains "relevant" under the Guidelines regime.
The decision to sweep unadjudicated offenses, and other criminal acts outside
of the offense of conviction into the sentencing inquiry has provoked an enormous
and almost uniformly hostile body of literature. See, e.g., Gerald W. Heaney, No
End to Disparity, 28 AX. CRIM. L. REV. 161, 185-225 (1991); Elizabeth T. Lear, Is
Conviction Irrelevant?, 40 UCLA L. REV. 1179 (1993); Boyce F. Martin, The Cor-
BROOKLYN LAW REVIEW
[Vol. 60: 725
bases punishment on a defendant's actual conduct as opposed
to the offense of conviction. The Guidelines sweep a variety of
factors into the sentencing inquiry, including criminal offenses
for which no conviction has been obtained. Under the Guidelines, therefore, prosecutorial charging decisions and even
verdicts of acquittal after jury trial may have little impact at
sentencing.
Long before the adoption of the Guidelines, courts bent on
rationalizing the real-offense regime devised a convenient yet
dangerous fiction in the form of the "punishment-enhancement" distinction. According to this theory, a sentence enhancement does not constitute punishment. Thus, challenges
to the use of nonconviction offenses 2 to increase sentence
length are easily answered: a defendant is not punished for the
unadjudicated offense; the sentence for the offense of conviction is merely enhanced on the basis of the extraneous criminal conduct.
The distinction is particularly convenient in the double
jeopardy setting. According to the Supreme Court's "favorite
saying,"3 double jeopardy "protects against a second prosecution for the same offense after acquittal[,] ... a second prosecution for the same offense after conviction[, and] ... against
multiple punishments for the same offense." Because a defendant whose sentence is "enhanced" for an unadjudicated crime
is neither "punished" nor tried for that offense, the Double
Jeopardy Clause is not technically implicated.
Recent opinions in the courts of appeals have reopened the
punishment-enhancement debate in the double jeopardy arena.
nerstone Has No Foundation:Relevant Conduct in Sentencing and the Requirements
of Due Process, 3 SETON HALL CONST. L.J. 25 (1993); Michael Tonry, Salvaging the
Sentencing Guidelines in Seven Easy Steps, 4 FED. SENT. REP. 355 (1992); David
Yellen, Illusion, Illogic and Injustice: Real-Offense Sentencing and the Federal Sentencing Guidelines, 78 MINN. L. REV. 403, 433-54 (1993).
2 The terms "nonconviction offense," "nonconviction crime," "unadjudicated offense," and "unadjudicated crime" are used interchangeably to refer to conduct that
has never been the subject of a conviction but is defined as criminal by statute.
"Acquittal conduct" and "acquittal offense" refer to crimes of which the defendant
has been acquitted.
' George C. Thomas, An Elegant Theory of Double Jeopardy, 1988 U. ILL. L.
REv. 827, 830.
" North Carolina v. Pearce, 395 U.S. 711, 717 (1969). The Double Jeopardy
Clause states in relevant part: "nor shall any person subject for the same offence
to be twice put in jeopardy of life or limb." U.S. CONST. amend. V.
1994]
DOUBLE JEOPARDY
In United States v. McCormick,5 the Second Circuit joined the
Tenth Circuit6 in holding that the Guidelines' use of unadjudicated crimes to increase sentence length constitutes punishment and that subsequent attempts to formally prosecute
those offenses are barred by the Double Jeopardy Clause. Interestingly, both courts found the punishment question reasonably straightforward, even though the punishment-enhancement fiction forms one of the central premises of twentiethcentury sentencing jurisprudence.
The Second Circuit's break with tradition in McCormick
provides an opportunity to reevaluate the punishment-enhancement distinction in the context of the Federal Sentencing
Guidelines. The punishment-enhancement fiction was originally embraced to repulse double jeopardy challenges to early
recidivist statutes.7 Defendants claimed that the harsher penalties imposed as a result of a prior conviction constituted a
second punishment for the prior offense. The Supreme Court
rejected this theory, reasoning that the existence of the previous conviction made the second transgression of the law more
serious.'
Over time this innocent pronouncement was extended well
beyond the prior-conviction or double jeopardy sphere. In response to relentless attacks on the Guidelines' real-offense
regime, the federal courts have repeatedly invoked the punishment-enhancement distinction, buttressing the position by
analogizing unadjudicated crimes to prior convictions. According to the courts, the use of prior convictions and the use of
nonconviction offenses at sentencing rest upon the same constitutional theory. Thus, if one rejects the punishment-enhancement rationale, one must similarly reject the use of prior
992 F.2d 437 (2d Cir. 1993).
See United States v. Koonce, 945 F.2d 1145 (10th Cir. 1991), cert. denied,
112 S. Ct. 1695 and 112 S. Ct. 1705 (1992).
See, e.g., McDonald v. Massachusetts, 180 U.S. 311, 312 (1901) ("The statute
under which [defendant's sentence] was rendered is aimed at habitual criminals;
and simply imposes a heavy penalty upon conviction of a felony . . . by one who
had been twice convicted."); Chenowith v. Commonwealth, 12 S.W. 585, 585 (Ky.
1889) ("The increased punishment is not for the former offenses; but the previous
convictions merely aggravate the last offense, and add to its punishment."); Kelley
v. People, 4 N.E. 644, 645 (Ill. 1886) ("There is no trial twice for the same offense, but twice for two crimes committed at different times.").
' McDonald, 180 U.S. at 312.
6
BROOKLYN LAW REVIEW
[Vol. 60:725
convictions to justify more severe sentences?
This Article takes issue with the "prior-conviction analogy," rejecting the notion that prior convictions and unadjudicated crimes are sufficiently similar to warrant equal treatment under the Constitution. Part I of this Article uses
McCormick to illustrate the subsequent-prosecution dilemma
under the Guidelines, describes the Second Circuit's solution,
and discusses the traditional response to the problem. Part II
questions the prior-conviction analogy, concluding that priorconviction enhancements are entirely consistent with the history and underlying values of the Double Jeopardy Clause. The
same cannot be said for unadjudicated-offense enhancements,
which strike at the heart of the double jeopardy protection,
wholly offending the central purposes of the Clause. Thus, Part
II concludes that the Second Circuit was correct in holding
that relevant conduct enhancements ° under the Guidelines
comprise constitutionally cognizable punishment.
Part III recognizes, however, that accurately characterizing a sentence enhancement as punishment does not automatically catapult the subsequent prosecution into a double jeopardy violation. Traditional double jeopardy jurisprudence does
not conform gracefully to the sentencing context. In fact, a
close analysis of the Second Circuit's approach suggests that
the lack of a conviction for the relevant conduct offense in the
original proceeding renders the Double Jeopardy Clause incompetent to protect its central purposes.
Part IV concludes that the Second Circuit chose the wrong
constitutional remedy. The Double Jeopardy Clause is not
designed to address the subsequent-prosecution problem in the
sentencing context. It is the Due Process Clause that forbids
enhancement in the absence of conviction in the first place.
Part IV therefore urges the courts and the Federal Sentencing
Commission to abandon the real-offense model once and for
all."
' See, e.g., United States v. Carey, 943 F.2d 44 (11th Cir. 1991), cert. denied,
112 S. Ct. 1676 (1992); United States v. Mack, 938 F.2d 678 (6th Cir. 1991).
" The relevant conduct provision, see U.S.S.G. § 1B1.3, is discussed in detail
infra at notes at 16-20 and accompanying text.
" This Article recognizes that abandoning the real-offense regime may eliminate a central difficulty with the Guidelines, but will create a whole new series of
issues that will need to be addressed. The major hurdle in designing a conviction-
1994]
DOUBLE JEOPARDY
I. THE SUBSEQUENT-PROSECUTION PROBLEM
A. A GuidelinesPrimer
For the uninitiated, the following discussion provides an
abbreviated explanation of the operation of the Federal Sentencing Guidelines.12 Inside the back cover of the FederalSentencing Guidelines Manual lies a deceptively simple-looking
numerical grid with "Criminal History" categories on one axis
and "Offense Levels" on the other." To arrive at the appropriate Guidelines sentence, one merely calculates the defendant's
criminal history score and offense level, locates the point at
which the two scores intersect, and reads off the applicable
range. The difficulty, of course, lies with arriving at the appropriate scores in the first instance.
To assure uniformity among "similarly situated offenders,"
the Guidelines embrace a real-offense model, 4 basing punishment decisions upon the defendant's actual conduct as opposed
to simply the offense of conviction. The Guidelines therefore
offense system will be to determine the type of information that should be used to
differentiate between offenders convicted of the same crime. The large number of
overlapping offenses set forth in the federal criminal code may make inevitable the
consideration of some conduct prohibited by a separate statute. See infra notes
138-40 and accompanying text. Delineating the extent to which "real factors," as
opposed to unadjudicated offenses, should be used at sentencing is beyond the
scope of this Article. For thoughtful approaches to this problem, see Reitz, supra
note 1, at 565-71; Yellen, supra note 1, at 454-60.
12 This description is intentionally oversimplified and is designed to allow a
reader unfamiliar with the intricacies of Guidelines sentencing to evaluate the
double jeopardy difficulties inspired by the subsequent prosecution of relevant conduct offenses.
13 See Marc Miller, True Grid: Revealing Sentencing Policy, 25 U.C. DAVIS L.
REV. 587, 599-600 (1992) (suggesting that a simpler, more informative grid is both
possible and preferable to the 258-box, 43-level version adopted by the Commission).
14 It was through the real-offense mechanism that the Commission sought to
prevent prosecutors from "circumventing" the Guidelines by charge bargaining. See
Stephen J. Schulhofer & Ilene H. Nagel, Negotiated Pleas Under the Federal Sentencing Guidelines: The First Fifteen Months, 27 AM. CRiM. L. REV. 231, 234
(1989). By insisting that an offender be punished for all his criminal acts without
regard to whether he has been charged or convicted of these offenses, the Guidelines theoretically eliminate the impact of prosecutorial decisions to drop charges.
" The explanation for the choice to embrace a real- as opposed to convictionoffense scheme is set forth in Stephen Breyer, The Federal Sentencing Guidelines
and the Key Compromises Upon Which They Rest, 17 HOFSTRA L. REV. 1, 8-12
(1988) and Ilene H. Nagel, Structuring Sentencing Discretion: The New Federal
BROOKLYN LAW REVIEW
[Vol. 60: 725
require a court to account for certain related crimes in the
offense-level computation, regardless of whether such offenses
have been the subject of a conviction. The relevant conduct
provision" constitutes the major, though not exclusive, means
by which unadjudicated offenses are included in the sentencing
calculus. 7 Section 1B1.3 extends the sentencing inquiry to all
acts jointly or individually undertaken by the defendant before,
Sentencing Guidelines, 80 J. CRIM. L. & CRIMINOLOGY 883, 925-26 (1990).
16 The
relevant conduct provision, crowned the "cornerstone" of the Guidelines
by former Commission Chairman William Wilkins, see William W. Wilkins Jr. &
John R. Steer, Relevant Conduct: The Cornerstone of the Federal Sentencing Guidelines, 41 S.C. L. REV. 495, 525 (1990), provides in pertinent part:
(a) Chapter Two (Offense Conduct) and Three (Adjustments). Unless
otherwise specified, (i) the base offense level where the guideline
specifies more than one base offense level, (ii) specific offense characteristics and (iii) cross references in Chapter Two, and (iv) adjustments in Chapter Three, shall be determined on the basis of the following:
(1) (A) all acts and omissions committed, aided, abetted, counseled,
commanded, induced, procured, or willfully caused by the
defendant; and
(B) in the case of a jointly undertaken criminal activity (a
criminal plan, scheme, endeavor, or enterprise undertaken
by the defendant in concert with others, whether or not
charged as conspiracy), all reasonably foreseeable acts and
omissions of others in furtherance of the jointly undertaken
criminal activity, that occurred during the commission of
the offense of conviction, in preparation for that offense, or
in the course of attempting to avoid detection or responsibility for that offense;
(2) solely with respect to offenses of a character for which §
3D1.2(d) would require grouping of multiple counts, all acts and
omissions described in subdivisions (1)(A) and (1)(B) above that
were part of the same course of conduct or common scheme or
plan as the offense of conviction;
(3) all harm that resulted from the acts and omissions specified in
subsections (a)(1) and (a)(2) above, and all harm that was the
object of such acts and omissions; ....
U.S.S.G. § 1B1.3(a).
'1 Unadjudicated crimes may also form the basis for a specific offense characteristic, see, e.g., U.S.S.G. § 2D1.1(b)(1) (requiring a two-level increase if a dangerous weapon was possessed during a drug offense, even though 21 U.S.C. §
924(c) prohibits use of a gun during the commission of a drug offense), an obstruction adjustment, see, e.g., U.S.S.G. § 3C1.1 comment. n.3(b) (requiring a two-level
increase if defendant obstructed the prosecution by committing perjury, even
though perjury is a separately punishable crime by statute), or departure. See, e.g.,
United States v. Kikumura, 918 F.2d 1084 (3d Cir. 1990) (allowing departure on
the basis of defendant's uncharged terrorist activities).
1994]
DOUBLE JEOPARDY
during or after the conviction offense,' 8 to acts that were "part
of the common course of conduct or common scheme or plan" as
the conviction offense," as well as to any harm resulting from
the preceding acts.2"
Thus, to determine a defendant's offense level, one begins
with the statutory offense of conviction.' Appendix A of the
Sentencing Guidelines provides a handy list of federal offenses
complete with the corresponding Guidelines provisions. After
locating the applicable guideline, the sentencer uses the conviction-offense conduct and all other acts deemed relevant under
section 1B1.3 to calculate a base offense level. This calculation
may then be adjusted for other factors, such as the defendant's
role in the offense22 or her acceptance of responsibility.
After the offense level is calculated, the sentencer moves
on to determine the offender's criminal history category.' The
criminal history score is calculated on the basis of criminal
convictions reaching back as far as fifteen years.25 Once the
criminal history category is obtained, it is plotted against the
offense level to yield a "sentencing range." At this point the
judge may impose a prison term within the range, or "depart"
if "there exists an aggravating or mitigating circumstance...
not adequately taken into consideration by the Sentencing
Commission,"" or if the government has moved for a depar18 U.S.S.G. § 1B1.3(a)(1)(b).
19 U.S.S.G. § 1B1.3(a)(2).
20 U.S.S.G. § 1B1.3(a)(3).
21 The Guidelines usually are
referred to as a "modified" real-offense system
because to some degree they hinge punishment on the offense of conviction and
because the Guidelines real-offense regime encompasses a narrower range of conduct than the pre-Guidelines rehabilitative model. See Nagel, supra note 15, at
926-27; Wilkins Jr. & Steer, supra note 16, at 497.
22 See, e.g., U.S.S.G. § 3B1.1(a) (requiring a four-level increase in offense level
"[i]f the defendant was an organizer or leader of criminal activity that involved
five or more participants or was otherwise extensive").
' See U.S.S.G. § 3El.1(a), 1.1(b) (allowing a two-level decrease in offense level
if "defendant clearly demonstrates acceptance of responsibility for his offense," and
an additional one-level decrease if defendant affirmatively assisted the investigation or prosecution of his offense).
24 See generally U.S.S.G. Ch.4.
See U.S.S.G. § 4A1.2(e). In limited instances where a defendant's criminalhistory score seriously underrepresents the true extent of his past criminal behavior, § 4A1.3 authorizes a departure from the applicable criminal history category.
26 18 U.S.C. § 3553(b) (1988). The Commission adopted policy statements regarding the propriety of using certain types of conduct or offender characteristics
as the basis for departure. See U.S.S.G. §§ 5H.1-12, 5K2.0-16. For more in-depth
BROOKLYN LAW REVIEW
[Vol. 60: 725
ture to reward a defendant's "substantial assistance."27
The impact of facts determined at sentencing can be dramatic. For example, in United States v. Rivera-Lopez,2" the
defendant was convicted of possession with intent to distribute
two kilograms of cocaine, but was acquitted by the same jury
of a second drug-possession county At sentencing, the district
court found under section 1B1.3(a)(2) that the acquittal conduct formed part of the "same course of conduct" as the crime
of conviction. Consequently, the court added the two kilograms
possessed as part of the conviction offense to the three kilograms attributable to the acquittal conduct to arrive at
defendant's base offense level."0 The inclusion of the acquittal
conduct increased the defendant's maximum sentencing expoinformation regarding departures in practice and criticism of the Commission's
attitudes toward departure innovations, see generally Karin Bornstein, 5K2.0 Departures for 5H Individual Characteristics:A Backdoor Out of the Federal Sentencing Guidelines, 24 COLUM. HUM. RTS. L. REV. 135 (1993); Daniel J. Freed, Federal
Sentencing in the Wake of the Guidelines: Unacceptable Limits on the Discretion of
Sentencers, 101 YALE L.J. 1681, 1715 (1991) (discussing the Commission's hostility
to individual factors as a basis for judicial departures); Aaron J. Rappaport, Reviewing Departures: The First Circuit's New Rivera Rule, 6 FED. SENT. REP. 43
(1993); Jean H. Shuttleworth, Childhood Abuse as a Mitigating Factor in Federal
Sentencing: The Ninth Circuit Versus the United States Sentencing Commission, 46
VAND. L. REV. 1333 (1993); Kirk D. Houser, Comment, Downward Departures: The
Lower Envelope of the Federal Sentencing Guidelines, 31 DUQ. L. REV. 361 (1993).
27 See U.S.S.G. § 5K1.1 (permitting a court to depart from even mandatory
minimum sentences upon a motion from the government stating that defendant
has provided substantial assistance to the investigation or prosecution of another
person). Critics have identified a number of problems with a government motion
requirement. First, the provision encourages disparity among similar offenders
because prosecutors do not use identical standards to determine when and whether
to move for departure. See Note, Prosecutorial Discretion and Substantial Assistance: The Power and Authority of Judicial Review, 15 CAMPBELL L. REv. 263, 273
n.72 (1993). Second, § 5K1.1 motions create disparity among co-defendants. Professor Freed cites the examples in drug cases, where minor offenders such as drug
couriers receive heavy sentences because they have little assistance to offer, while
kingpins receive lighter sentences after divulging information. See Freed, supra
note 26, at 1705; see also Antoinette M. Tease, Downward Departures For Substantial Assistance: A Proposal for Reducing Sentencing DisparitiesAmong Co-Defendants, 53 MONT. L. REV. 75 (1992).
The Supreme Court settled one thorny issue in this area in United States v.
Wade, 112 S. Ct. 1840 (1992), by holding that the statutory requirement of a
government motion before departure under § 5K1.1 could be overcome by a "substantial threshold showing" that the prosecutor acted in bad faith in withholding
the motion. Id. at 1844.
28 928 F.2d 372 (11th Cir. 1991).
29 Id. at 372.
0 Id. at 372-73.
1994]
DOUBLE JEOPARDY
sure from 97 months to 151 months.
Even though facts found at sentencing are of great importance to the defendant, sentencing hearings remain informal
events conducted with little regard for procedural fairness or
accuracy."' Sentencing facts need only be proven by a preponderance of the evidence." Neither the Federal Rules of Evidence33 nor the Confrontation Clause apply. 4 And although
defendants technically have a due process right to be sentenced only upon reliable information,35 offenders are routine-
3 The literature attacking the procedures at sentencing is extensive. See, e.g.,
Gerald W. Heaney, The Reality of Guidelines Sentencing: No End to Disparity, 28
AM. Cmi. L. REV. 161, 208-25 (1991); Susan N. Herman, The Tail That Wagged
the Dog: Bifurcated Fact-FindingUnder the Federal Sentencing Guidelines and the
Limits of Due Process, 66 S. CAL. L. REV. 289, 323-43 (1992); Deborah Young,
Fact-Findingat Federal Sentencing: Why the Guidelines Should Meet the Rules, 79
CORNELL L. REV. 299 (1994); Joseph P. Sargent, Comment, The Standard of Proof
Under the Federal Sentencing Guidelines: Raising the Standard to Beyond a Reasonable Doubt, 28 WAKE FOREST L. REV. 463 (1993).
32 See, e.g., United States v. Nyhuis, 8 F.3d 731, 744 (11th Cir. 1994), cert.
denied, 63 U.S.L.W. 3257 (U.S. Oct. 4, 1994); United States v. Restrepo, 903 F.2d
648 (9th Cir. 1990), cert. denied, 112 S. Ct. 1564 (1992); United States v. Dawn,
897 F.2d 1444, 1449-50 (8th Cir. 1990), cert. denied, 498 U.S. 960 (1990). A number of commentators have criticized the preponderance standard at sentencing as
inadequate. See, e.g., Richard Husseini, Note, The Federal Sentencing Guidelines:
Adopting Clear and Convincing Evidence as the Burden of Proof, 57 U. CHI. L.
REV. 1387 (1990); Sargent, supra note 31, at 471-77.
' See United States v. Prescott, 920 F.2d 139, 144 (2d Cir. 1990) (allowing use
of hearsay statements at sentencing); United States v. Jarrett, 705 F.2d 198, 208
(7th Cir. 1983) (allowing suspension of best evidence rule at sentencing), cert.
denied, 465 U.S. 1004 (1984). Several commentators have offered excellent arguments for the application of the Federal Rules of Evidence at the sentencing hearing. See generally Edward R. Becker & Aviva Orenstein, The Federal Rules of
Evidence After Sixteen Years-The Effect of "Plain Meaning" Jurisprudence, the
Need for an Advisory Committee on the Rules of Evidence, and Suggestions for
Selective Revision of the Rules, 60 GEO. WASH. L. REV. 857, 885-91 (1992);
Randolph N. Jonakait, Insuring Reliable Fact Finding in Guidelines Sentencing:
Why Not Real Evidence Rules?, 22 CAP. U. L. REV. 31 (1993); Young, supra note
31, at 301.
"' See, e.g. United States v. Kikumura, 918 F.2d 1084, 1102 (3d Cir. 1990)
(refusing to "superimpose the jurisprudence of the confrontation clause upon the
sentencing phase"). For a thoughtful argument that the Confrontation Clause
should apply at sentencing, see Note, An Argument for Confrontation Under the
Federal Sentencing Guidelines, 105 HARV. L. REV. 1880, 1888-98 (1992).
" See Townsend v. Burke, 334 U.S. 736, 740-41 (1948); David McGovern et al.,
Project, Twenty-Third Annual Review of Criminal Procedure: United States Supreme Court and Court of Appeals, 1992-1993, 82 GEO. L.J. 597, 1188 (1994) ("Due
process forbids the judge from relying on materially false or unreliable information
BROOKLYN LAW REVIEW
[Vol. 60: 725
ly condemned to spend years behind bars on the basis of double or triple hearsay and questionable drug quantity estimates."
B. The Subsequent-ProsecutionProblem Under the Federal
Sentencing Guidelines
United States v. McCormick37 provides a simple example
of the subsequent-prosecution problem under the Federal Sentencing Guidelines. McCormick was indicted in the District of
Connecticut for bank fraud and related crimes involving losses
of approximately $75,000.38 Some three months later,
McCormick was again indicted, but this time in the District of
Vermont. According to the Vermont indictment, McCormick
had engaged in a series of successful and unsuccessful attempts to defraud banks and financing companies in Vermont,
as well as obstruction of justice during the investigation of
those offenses. The estimated loss attributable to the Vermont
crimes exceeded $4,000,000."9 The Connecticut prosecution
proceeded to trial first, and McCormick was convicted on all
counts.40
When the district court calculated McCormick's Guidelines'
sentence, it properly considered both the fraudulent acts for
" See, e.g., United States v. Easterling, 921 F.2d 1073, 1077-78 (10th Cir.
1990) (To determine the relevant drug quantity for sentencing, the district court
relied on the testimony of a government agent who claimed that two unidentified
informants estimated purchasing "approximately two pounds each" of methamphetamine from the defendant. The court added four pounds to the .08 grams of which
defendant had been convicted of distributing to arrive at the base offense level.),
cert. denied, 500 U.S. 937 (1991).
The use of drug quantities to determine punishment levels has proven extremely controversial. See Albert W. Alschuler, The Failure of the Sentencing
Guidelines: A Plea for Less Aggregation, 58 U. CHI. L. REV. 901, 921-22 (1991);
Eric P. Berlin, Comment, The Federal Sentencing Guidelines' Failure to Eliminate
Sentencing Disparity: Governmental Manipulations Before Arrest, 1993 WiS. L. REV.
187; Thomas J. Meir, Comment, A Proposal to Resolve the Interpretation of "Mixture or Substance" Under the Federal Sentencing Guidelines, 84 J. CRIM. L. &
CRIMINOLOGY 377 (1993); Douglas J. Quivey, Note, Market-Oriented Approach to
Determining Drug Quantity Under the Federal Sentencing Guidelines, 1993 U. ILL.
L. REV. 653.
37 992 F.2d 437 (2d Cir. 1993).
"8Id. at 438.
39 Id.
40
Id.
1994]
DOUBLE JEOPARDY
which McCormick was convicted and those offenses set forth in
the pending Vermont indictment.4' Section 2F1.1(b)(1), the
applicable guideline for bank fraud, keys punishment to the
amount of money stolen.4' Thus, the inclusion of the Vermont
frauds generated a seven-level increase in McCormick's base
offense level and extended his maximum sentencing exposure
from twenty-one to forty-six months.43
Id. McCormick's counsel objected to the inclusion of the Vermont conduct,
then the subject of a pending indictment. United States v. McCormick, 798 F.
Supp. 203, 206 (D. Vt. 1992), aff'd, 992 F.2d 437 (2d Cir. 1993). In a letter to the
court, the government responded by claiming that "[a] sentence based in part on
the Vermont conduct will have the effect of barring further prosecution on the
Vermont charges. The United States Attorney's office for the District of Vermont
would have no objection to such a sentence if the court sentences within the
Guidelines range appropriate to the combined total losses of all the defendant's
known fraudulent schemes." Joint Appendix, at A134 (quoting letter of Oct. 30,
1991, to Judge Nevas from Assistant U.S. Attorney Hutchison of the District of
Connecticut). At the sentencing hearing, the government "clarified" the position of
the District of Vermont, pointing out that "it is unclear to them that they are
barred from further prosecution whether or not-if the Court sentences on the
Vermont fraud. It would appear that they would not be likely to pursue their
charges, but I don't want to represent to the Court that that conclusion has been
reached by them at this point." Joint Appendix, at A137 (quoting transcript of sentencing hearing before Judge Nevas on November 5, 1991). After exploring the
ramifications of including the Vermont losses, the district court adopted the
government's position. McCormick, 798 F. Supp. at 206.
42 Under U.S.S.G. § 2F1.1(b), for example, a loss of under $2000 yields a base
offense level of 6, while losses between $40,000 and $70,000 provoke a five-level
increase, yielding a base offense level of 11.
' To calculate McCormick's sentence, one must first turn to U.S.S.G. § 2F1.1,
which keys the base offense level to the amount of money stolen. But remember
the relevant conduct provision requires the sentencer to consider acts outside the
offense of conviction. See supra note 16 and accompanying text. Consequently, §
1B1.3 should be consulted to determine the scope of activity to be included in the
inquiry. In this case, § 1B1.3(a)(2) provides the proper starting point. This section
incorporates yet another Guidelines provision-§ 3D1.2(d)-which treats offenses as
a "group" if they rely on either drug or money quantities to determine punishment
levels. Since § 2F1.1 is keyed to the amount of money stolen, § 1B1.3(a)(2) requires the court to consider monies obtained as "part of the same course of conduct or common scheme or plan as the offense of conviction." U.S.S.G. §
1B1.3(a)(2). The similarities in the schemes employed to defraud the Vermont and
Connecticut financial institutions justify the inclusion of the losses attributable to
the Vermont frauds in the base offense-level calculation. Adding the $75,000 in
losses stemming from McCormick's actions in Connecticut to the Vermont losses
yields a total of $ 4,180,000. See 798 F. Supp. at 210 n.7. Under § 2F1.1(b), this
amount warrants a base offense level of 19. In addition, § 2Fl.l(b)(2) requires a
two-level increase for more than minimal planning, resulting in an offense level of
21. Because McCormick had no prior convictions, he fell into criminal history category I. His sentencing ranges was therefore 37 to 46 months. In the absence of
41
BROOKLYN LAW REVIEW
[Vol. 60: 725
Most of the conduct contained in the Vermont indictment
was accounted for in McCormick's Connecticut sentence, yet
the U.S. Attorney's office for the District of Vermont declined
to forego the second prosecution, thus provoking what I term
the "subsequent-prosecution dilemma." After an unsuccessful
appeal challenging the inclusion of the pending Vermont offenses as relevant conduct, McCormick filed a motion to dismiss the Vermont indictment on double jeopardy grounds. The
district court granted McCormick's motion as to all counts
alleging conduct that was actually used to increase the length
of McCormick's Connecticut sentence."
1. The Second Circuit Solution
The Second Circuit affirmed, agreeing that the prosecution
of the Vermont offenses would violate the "multiple punishment" prong of the Double Jeopardy Clause.45 To reach this
conclusion, the court relied on a three-part analysis developed
by the Tenth Circuit in United States v. Koonce.4 Pursuant to
the Koonce approach, the court considered the following issues:
(1) whether increasing the offense level for the conduct at issue
in the initial sentencing proceeding constituted punishment
within the meaning of the Double Jeopardy Clause; (2) whether Congress intended to allow a second punishment for that
conduct; and (3) whether the imposition of concurrent rather
than consecutive sentences would avoid the double jeopardy
problem.47
The Second Circuit found the punishment inquiry
the Vermont offenses, McCormick's sentencing range would have been 15 to 21
months. McCormick was sentenced to 46 months of imprisonment. See 798 F.
Supp. at 206.
41 Most of the acts set forth in the Vermont indictment were included in the
Connecticut sentencing memorandum, but not all of the acts had resulted in an
actual loss to the victim and a corresponding increase in base offense level under
U.S.S.G. § 2F1.1. Because those offenses did not technically increase McCormick's
sentence, the district court concluded that no multiple punishment issue was presented under the Double Jeopardy Clause. McCormick, 798 F.2d at 210-12.
" McCormick, 992 F.2d 437 (2d Cir. 1993). For a brief discussion of the multiple punishment prong of the Double Jeopardy Clause, see infra notes 120-29 and
accompanying text.
46 945 F.2d 1145 (10th Cir. 1991), cert. denied, 112 S. Ct. 1695 and 112 S. Ct.
1705 (1992).
' McCormick, 992 F.2d at 439.
19941
DOUBLE JEOPARDY
"straight-forward," concluding that
McCormick was punished for the Vermont conduct that was taken
into account by the Connecticut court when it determined the
amount of loss for which McCormick was responsible. The government requested that the court take this conduct into account, the
court explicitly stated that it was taking the conduct into account,
and the ultimate 'sentence reflects part of McCormick's Vermont con-
duct. 8
The McCormick court divined no intent on the part of Congress to authorize a second punishment for the conduct at
issue. The court pointed out that the Guidelines were enacted
to alleviate disparity among sentences "'imposed for similar
criminal offenses committed by similar offenders."' 9 Permitting a second punishment for the relevant conduct offenses
would allow the government to evade this goal through piecemeal prosecution."
Lastly, the McCormick court considered whether concurrent sentences would avoid the double jeopardy problem. According to Ball v. United States,5 an additional conviction
produces adverse collateral consequences, such as enhanced
punishment pursuant to recidivist statutes, "possible delays in
parole eligibility, and [general] societal stigma,"52 which cannot be eliminated through concurrent sentencing. Consequently, the Second Circuit rejected the notion that concurrent sentences would cure the double jeopardy defect in McCormick."'
48 Id. The court of appeals agreed with the district court that conduct that had
not actually contributed to an increase in McCormick's base offense level in the
original proceeding had not been punished and was, therefore, outside the scope of
the double jeopardy bar. Id. at 439-40.
" Id. at 440 (quoting U.S.S.G. Ch.1, Pt.A(3), p.s.).
'0 Id. at 440-41.
5, 470 U.S. 856 (1985).
82 McCormick, 798 F. Supp. 203, 209.
McCormick, 992 F.2d at 441. The decision drew a stinging dissent from
Judge Mahoney, who complained that the majority had ignored the operation of
Guidelines § 5G1.3(b), which was drafted specifically to deal with the instant problem. Id. at 443 n.2 (Mahoney, J., concurring & dissenting in part). Section 5G1.3
provides in pertinent part: "If ... the undischarged term of imprisonment resulted
from offense(s) that have been fully taken into account in the determination of the
offense level for the instant offense, the sentence for the instant offense shall be
imposed to run concurrently to the undischarged term of imprisonment." Id. at
443. According to commentary quoted by the dissent the "guideline is intended to
result in an appropriate incremental punishment for the instant offense that most
BROOKLYN LAW REVIEW
[Vol. 60: 725
2. The TraditionalApproach
Although the Second Circuit sought to downplay the renegade nature of the decision,54 the McCormick approach arguably runs counter to nearly a century of sentencing jurisprudence. Courts typically have avoided the reach of the Double
Jeopardy Clause by refusing to acknowledge that a sentence
enhancement based upon the existence of an unadjudicated offense constitutes punishment for that offense. This punishment-enhancement distinction derives from the early Supreme
5 5 in which the vaCourt case of McDonald v. Massachusetts,
nearly approximates the sentence that would have been imposed had all the sentences been imposed at the same time." Id. (quoting U.S.S.G. § 5G1.3 comment.
(backg'd)). Thus, according to the dissent, no issue regarding congressional intent
to allow double punishment ever arises because § 5G1.3 "recognizes and explicitly
addresses the sentencing problem posed by successive prosecutions in which an
overall course of conduct is segmented into separate criminal charges, but § 1B1.3
requires that the entire relevant conduct nonetheless be addressed in sentencing at
both proceedings." 992 F.2d at 443.
Judge Mahoney technically filed an opinion concurring in part and dissenting
in part. Id. at 442. The opinion dissents from the majority's double jeopardy holding, concurring only with the decision not to block the prosecution of the Vermont
acts that had no effect on the Connecticut sentence. See id.
"' In McCormick, Judge Oakes dedicated only two paragraphs to distinguish
cases upholding "conviction and punishment for conduct that was previously used
to enhance a defendant's sentence for other conduct." 992 F.2d at 441. Citing preand post-Guidelines cases that arguably conflict with the McCormick holding in a
footnote, Id. at 441 n.5, he explained that "[tihe critical distinction between these
other cases and McCormick's case is found in our analysis of congressional intent.
If Congress intends to allow the same conduct to be used to enhance a sentence
and to serve as the basis for a separate prosecution, the Double Jeopardy Clause
does not stand in the way." Id. at 442.
In United States v. Koonce, the Tenth Circuit also distinguished several
Guidelines cases that had allowed a subsequent conviction and punishment for
conduct initially used for enhancement purposes. 945 F.2d 1145, 1152 (10th Cir.
1991) (discussing United States v. Mack, 938 F.2d 678 (6th Cir. 1991)), cert. denied, 112 S. Ct. 1695 and 112 S. Ct. 1705 (1992); United States v. Williams, 935
F.2d 1531 (8th Cir. 1991), cert. denied, 112 S. Ct. 1189 (1992); United States v.
Garcia, 919 F.2d 881 (3d Cir. 1990)). The distinguishing feature seemed to be that
these cases did not involve relevant conduct enhancements. Rather, each case
involved an enhancement made pursuant to the Guidelines' Chapter Three adjustments, which take into account conduct such as obstruction of justice and failure
to appear at sentencing. Although the court did not explain why the point at
which the unadjudicated offense is accounted for in the Guidelines calculation
would make a difference, it implied that congressional intent to provide additional
punishment for these offenses might be different from that regarding relevant
conduct. Koonce, 945 F.2d at 1153.
5 180 U.S. 311 (1901).
1994]
DOUBLE JEOPARDY
lidity of an enhancement based upon a prior conviction was upheld against a double jeopardy challenge. In a terse decision,
the Court concluded that the prior-conviction enhancement did
not constitute a second punishment for the earlier offense;
rather, the existence of the former conviction amplified the
seriousness of the current offense, thus justifying a more extreme sentence.56
Sentencing decisions both before and after the adoption of
the Guidelines have relied on the McDonald punishment-enhancement rationale to defend the use of acquittal offenses and
" The McDonald holding probably came as no surprise to the legal community
of the day. Even prior to 1901, the state courts uniformly defended double jeopardy challenges to recidivist statutes on the theory that the prior transgression of
the law warranted a harsher penalty than would otherwise be required. See, e.g.,
Kelley v. People, 4 N.E. 644, 645 (Ill. 1886) ("There is no trial twice for the same
offense, but twice for two crimes committed at different times."); Chenowith v.
Commonwealth, 12 S.W. 585, 585 (Ky. 1889) ("The increased punishment is not for
the former offenses; but the previous convictions merely aggravate the last offense,
and add to its punishment."). This rationale has been repeatedly upheld in the
recidivist statute context. See Gryger v. Burke, 334 U.S. 728, 732 (1948).
McDonald acted as a springboard for other decisions rationalizing the realoffense regime. The McDonald punishment-enhancement theory dovetailed nicely
with the rehabilitative model embraced by the Court in Williams v. New York,
337 U.S. 241 (1949). Under the rehabilitative model, any information relevant to a
defendant's prospects for rehabilitation formed a legitimate object for inquiry at
sentencing. As a result, judges relied on uncharged criminal conduct as well as a
wide range of moral and social information about the defendant to devise the
appropriate sentence. For more information about the history of the rehabilitative
system, see generally FRANCIS A. ALLEN, THE DECLINE OF THE REHABILITATIVE
IDEAL (1981); Nagel, supra note 15, at 893-95; see also Reitz, supra note 1, at 52831.
It was not too long after Williams v. New York that the Court, in Williams v.
Oklahoma, 358 U.S. 576 (1969), upheld the use of a murder conviction to justify
the imposition of the death penalty for a related kidnapping charge. The double
jeopardy analysis in Williams v. Oklahoma was again short and to the point, noting only that Williams v. New York sanctioned the use of any information relevant
to defendant's "prospects for rehabilitation," id. at 584, and that the defendant was
not punished twice for the murder, only more severely for the kidnapping at issue.
Id. at 587.
Later Supreme Court decisions such as United States v. Grayson, 438 U.S. 41
(1978), and McMillan v. Pennsylvania, 477 U.S. 79 (1986), further contorted the
punishment-enhancement premise, extending it well beyond the realm of the priorconviction enhancement. Many scholars have attacked the faulty due process theory upon which Williams v. New York, McMillan and Grayson are premised. See
Herman, supra note 31, at 319-21, 325-37; Lear, supra note 1, at 1209-14. Rather
than rehash this argument, this Article reevaluates the original explanation for
the punishment-enhancement fiction and the extension of this reasoning beyond
the prior-conviction context.
BROOKLYN LAW REVIEW
[Vol. 60: 725
nonconviction crimes at sentencing. 7 For example, in a decision arguably inconsistent with McCormick,58 the Second Circuit invoked the punishment-enhancement fiction to reject a
double jeopardy challenge to an enhancement for acquittal
conduct. According to United States v. Rodriguez-Gonzalez,59
"the district court 'was not relying on facts disclosed at trial to
punish the defendant for the extraneous offense, but to justify
the heavier penalties for the offenses for which he was convicted."'6 0 In a dubious attempt to justify the fiction, the court
added: "Of course, from the point of view of the defendant
receiving added prison time because of the presence of a gun,
the distinction may be academic. But the analysis is crucial in
considering the double jeopardy argument."6'
Although Rodriguez-Gonzalez involved an acquittal enhancement, the majority of the courts of appeals have rejected
the McCormick holding in the subsequent-prosecution context
as well.62 For example, the Sixth Circuit rejected a challenge
to a subsequent conviction for failure to appear at sentencing,
noting that
"' See, e.g., United States v. Cruce, 21 F.3d 70 (5th Cir. 1994) (post-Guidelines
acquittal for nonconviction offense), cert. denied, 63 U.S.L.W. 3262 (U.S. Oct. 4,
1994); United States v. Juarez-Ortega, 866 F.2d 747, 748-49 (5th Cir. 1989) (postGuidelines enhancement for acquittal conduct); Sekou v. Blackburn, 796 F.2d 108,
111 (5th Cir. 1986) (pre-Guidelines enhancement for uncharged offense); United
States v. Sweig, 454 F.2d 181, 184 (2d Cir. 1972) (pre-Guidelines enhancement for
acquittal offense).
68 992 F.2d 437 (2d Cir. 1993).
'9 899 F.2d 177 (2d Cir.), cert. denied, 498 U.S. 844 (1990).
60 Id. at 181 (quoting United States v. Juarez-Ortega, 866 F.2d 747, 749 (5th
Cir. 1989) (per curiam)).
r1 Id.
62 See,
e.g., Williams v. Oklahoma, 358 U.S. at 584-86; Cruce, 21 F.3d at 74;
United States v. Carey, 943 F.2d 44, 46-47 (11th Cir. 1991), cert. denied, 112 S.
Ct. 1676 (1992); United States v. Mack, 938 F.2d 678, 681 (6th Cir. 1991); United
States v. Williams, 935 F.2d 1531, 1539 (8th Cir. 1991), cert. denied, 112 S. Ct.
1189 (1992); United States v. Garcia, 919 F.2d 881, 887 (3d Cir. 1990).
In Cruce, the Fifth Circuit confronted a subsequent-prosecution situation almost identical to that in McCormick. Like the defendant in McCormick, Cruce
argued that the Double Jeopardy Clause barred prosecution for an offense already
accounted for through relevant conduct at sentencing. The Fifth Circuit, however,
had no trouble rejecting the double jeopardy claim. According to the Cruce court,
defendant's argument "misconceive[d] the focal point of the double jeopardy protection against a second punishment as the conduct considered in setting the...
sentence, instead of the conviction offense that prompted the . . . sentence." 21
F.3d at 74. In other words, the defendant was not punished for the relevant conduct offense, he was punished for the offense of conviction.
19941
DOUBLE JEOPARDY
[The Guidelines] simply memorialize[ I a long standing practice of
trial judges. Prior convictions... are routinely used to determine
the quantity of sentence. If Appellant is correct, then any consideration [of extraneous offenses] is a second punishment and therefore
is a violation of the Double Jeopardy Clause.63
Similarly, the Eleventh Circuit defended the punishment-enhancement premise by explaining that an enhancement
is not considered "punishment" for that conduct in the double jeopardy context because the court is sentencing the defendant only for the
instant offense, which is considered more serious because of the
defendant's other criminal conduct. For this reason, the defendant's
prior criminal convictions also may be used to impose a harsher
sentence.64
II. PUNISHMENT, ENHANCEMENT AND THE PRIOR-CONVICTION
ANALOGY
The punishment-enhancement distinction forms one of the
mainstays of the real-offense regime, yet as the cases demonstrate, courts seldom subject the theoretical underpinnings of
the doctrine to critical analysis. Courts apparently assume that
the punishment-enhancement distinction applies with equal
force to enhancements for prior convictions and those involving
unadjudicated offenses. But a conviction obtained after a fair
trial, and an accusation made in the absence of even the most
basic constitutional protections, are not the same.65 To treat
them equally speaks volumes about the lackluster analysis to
which sentencing issues are characteristically subjected.
Although the existence of a prior conviction is instinctively
relevant to a subsequent punishment decision, the reasoning
set forth by the Supreme Court in McDonald and similar cases
explains only why we might want to increase punishment upon
proof of a second conviction; it does not explain why the Constitution allows us to do so. A close review of the McDonald
rationale demonstrates that the Court reasoned in reverse. The
Court assumed that if it acknowledged that the prior offense
Mack, 938 F.2d at 681 (emphasis added).
United States v. Carey, 943 F.2d 44, 46 n.4 (11th Cir. 1991) (emphasis added), cert. denied, 112 S. Ct. 1676 (1992).
" See supra notes 31-36 and accompanying text.
BROOKLYN LAW REVIEW
[Vol. 60: 725
was indeed punished by the harsher penalty at sentencing a
double jeopardy violation would automatically follow. Because
such a result was both intuitively and politically unpalatable,
the Court simply avoided the double jeopardy question altogether by refusing to accept the punishment description.6 6
Had the McDonald Court simply accepted the defendant's
punishment premise, the flaws in the prior-conviction analogy
would have been immediately apparent. Construing a priorconviction enhancement as punishment does not alter the
McDonald result-a second punishment is entirely consistent
with the structure and purposes of the Double Jeopardy
Clause. Punishment for nonconviction crimes, on the other
hand, is not authorized by the Double Jeopardy Clause, is
incompatible with the affirmative commands of other constitutional provisions,' and is wholly at odds with the basic values
of the double jeopardy protection.
The Double Jeopardy Clause was intended to be "declaratory of the law as it ... stood" and "to conform to universal
practice in Great Britain and this country."" As many scholars have noted, the scope of the double jeopardy protection was
no clearer in 1791 than it is today.6 9 History suggests, however, that punishment for prior convictions at sentencing was
presumed by the language of the Double Jeopardy Clause itself.
According to Professor Jay Sigler,
the law during the reign of Henry I provided that the punishment
upon a second conviction was death or mutilation for almost any offense. Similar provisions are to be found as far back as the laws of
Cnut and Ethelred. Only a few crimes were punishable by death or
mutilation on the conviction of the first offense.7"
Thus, the phrase "life or limb" used in the final version of the
McDonald v. Massachusetts, 180 U.S. 311, 312 (1901).
Real-offense sentencing allows the government to obtain punishment for a
criminal offense without adhering to the affirmative commands of the Fifth and
Sixth Amendments. See Lear, supra note 1, at 1221, 1223-37.
"8 Jay A. Sigler, A History of Double Jeopardy, 7 AM. J. LEG. HIST. 283, 298
(1963) (citing Note, 53 Nw. U. L. REv. 521 (1958)).
Id.; Thomas, supra note 3, at 827-28.
ToSigler, supra note 68, at 285 (quoting 1 JAMES F. STEPHEN, A HISTORY OF
THE CRIMINAL LAW OF ENGLAND 58-59 (1883) (emphasis added)). During this time,
"[aill crimes were punishable, instead, by . . . systems of compensation to be made
to the victims' relations or to the king or local lord." Id. at 285.
66
"
1994]
DOUBLE JEOPARDY
Fifth
Amendment "ha[d] a literal meaning in English histo71
ry; ;
it referred to the practice of punishing an offender more
severely upon proof of a second offense.7 2 This practice was
followed in the American colonies, where the death penalty
was applicable to a wide variety of first offenses, and also was
used to punish the repeat offender.7 3 Perhaps, then, the Double Jeopardy Clause, by including the "life or limb" qualification,7 4 can be read quite literally to authorize additional punishment for previously adjudicated offenses upon conviction of
a subsequent crime.7"
The legitimacy of exempting prior convictions from the
reach of the double jeopardy prohibition draws further support
71
Sigler, supra note 68, at 285.
2 For example, "in Ethelred's laws it
[was] said of an accused that upon a
second conviction 'let him be smitten so that his neck break.'" Sigler, supra note
68, at 285 (emphasis added).
Sigler, supra note 68, at 303.
See supra note 4.
7' Two difficulties with this analysis become immediately obvious. First, the
practice in ancient England and the American colonies is equally consistent with
the notion that the Constitution does not view harsher penalties upon a second
conviction as punishment for the original offense. Thus, the history can be read to
support the McDonald Court's reasoning, and that of a large number of early decisions upholding recidivist statutes. See, e.g., Carlesi v. New York, 233 U.S. 51
(1914) (upholding conviction under New York recidivist statute notwithstanding
executive pardon); Graham v. West Virginia, 224 U.S. 616 (1912) (upholding West
Virginia recidivist statute); Moore v. Missouri, 159 U.S. 673 (1895) (upholding
recidivist statute). My point, however, is not that the McDonald Court was disingenuous or incorrect; rather, I am attempting to demonstrate that the McDonald
Court's choice of words was completely irrelevant in the prior-conviction context. It
is only when that word choice is adapted to the nonconviction-offense scenario that
the punishment-enhancement designation becomes important.
Second, the Court has always treated the "life or limb" language as defining
the offenses to which the Double Jeopardy Clause should extend. Although a literal interpretation of the constitutional language would restrict double jeopardy
protection to cases involving capital or corporal punishment, the Supreme Court
long ago rejected this position. In Ex parte Lange, 85 U.S. (18 Wall.) 163, 170
(1873), the Court held that double jeopardy protection extended to all crimes, not
simply serious ones. See generally George C. Thomas III, A Modest Proposal to
Save the Double Jeopardy Clause, 69 WASH. U. L.Q. 195 (1991) (questioning
whether some of the definitional difficulties with the same offense language might
be avoided by limiting the scope of double jeopardy protection under the "life or
limb" formula). I do not dispute the logic of treating "life or limb" as defining the
class of cases to which the Double Jeopardy Clause should apply, but neither do I
suppose that the phrase must be restricted to one use or the other. Treating it as
enshrining a well-established exception to multiple punishment for prior convictions
does not preclude it from delineating the types of crimes to which the Clause
should extend.
BROOKLYN LAW REVIEW
[Vol. 60: 725
from the underlying values of the Double Jeopardy Clause.
Scholars have yet to agree on a single unifying theory of double jeopardy. 6 At base, however, the Double Jeopardy Clause
was designed to protect the individual and the "legal system itself."" From the defendant's perspective, Justice Black's oftquoted statement from United States v. Green eloquently sums
up the purposes of the Double Jeopardy Clause:
The underlying idea, one that is deeply ingrained in at least the
Anglo-American system of jurisprudence, is that the State with all
its resources and power should not be allowed to make repeated
attempts to convict an individual for an alleged offense thereby subjecting him to embarrassment, expense and ordeal and compelling
him to live in a continuing state of anxiety and insecurity, as well as
enhancing the possibility that even though innocent he may be
found guilty.78
Punishment for prior convictions at sentencing is completely consistent with these goals. Prior-conviction enhancements do not expose a defendant to embarrassment or expense
beyond that already incurred. There is no question regarding
the defendant's legal innocence. Her status has long been determined; the judgment has been entered. Using that conviction does not ask the defendant to defend the charge again, nor
does it give the government a chance to perfect its presentation
of the case.
Juxtapose this against punishment for an unadjudicated
offense. A defendant may be forced to respond to a single allegation numerous times. v9 And what Professor Friedland con-
7"See Thomas, supra note 3, at 828-34 (noting that "the proliferation of case
law and commentary has not produced a coherent theory to date," and identifying
three competing double jeopardy theories: "traditional analysis, the RehnquistO'Connor government oppression theory, and the verdict finality theory").
71 MARTIN L. FRIEDLAND, DOUBLE JEOPARDY 4 (1969).
7' Green v. United States, 355 U.S. 184, 187-88 (1957).
79Under prevailing double jeopardy jurisprudence, a single, criminal episode
may spawn a series of prosecutions for overlapping statutory violations. For example the sale of a controlled substance might provoke separate prosecutions for
conspiracy, use of a telephone in connection with a drug transaction, distribution
within 1000 feet of a school, and several other drug-related offenses. The quantity
of drugs attributable to the conviction offense would be determined pursuant to
the relevant conduct provision in each proceeding. See U.S.S.G. § 5G1.3(b) (discussed supra note 53 and accompanying text). Thus, although the defendant is
never formally charged, he may be forced to respond repeatedly to the same allegations.
1994]
DOUBLE JEOPARDY
tends is the "core of the problem"--the possibility that an innocent person may be convicted of a crime after repeated dry
runs by the govermnent--becomes a very real possibility
under the punishment-enhancement premise.
When the government alleges at sentencing that a defendant committed crimes in addition to the conviction offense,
the defendant is afforded an opportunity to contest the accusation and to present evidence on the point.8 As previously
pointed out, however, the sentencing hearing operates outside
the ambit of the Federal Rules of Evidence and the Confrontation Clause, and seldom demands more than merely proof by a
preponderance of the evidence to support government accusations.82 Extensive enhancements based upon double and triple
hearsay are common.83 Because the government's burden of
proof at sentencing is so slight, the defendant essentially faces
an affirmative-defense situation. The government need only
raise the issue while the defendant must create sufficient
doubt to undermine the court's acceptance of the government's
accusations.
The government may therefore receive an important preview of the defendant's evidence regarding a disputed issue
while disclosing very little of its own evidence. But because the
proceeding has been excluded from double jeopardy protection,
a finding in the defendant's favor affords the defendant no
relief. Having had the "dress rehearsal," the government may
simply prepare a more effective presentation for the next sentencing hearing or trial. This is especially troubling in the
subsequent-prosecution scenario because the defendant's testimony at sentencing or that of a witness will be admissible in a
later trial, whether or not the defendant elects to testify at the
subsequent proceeding.'
Of course, the traditional practice under the Guidelines
"
FRIEDLAND, supra note 77, at 4.
',
See U.S.S.G. § 6A1.3(a) (requiring that parties be "given an adequate oppor-
tunity to present information to the court regarding" disputed allegations at sen-
tencing).
See supra text accompanying notes 31-36.
See, e.g., United States v. Easterling, 921 F.2d 1073 (10th Cir. 1990), cert.
denied, 500 U.S. 937 (1991); see also Heaney, supra note 31, at 210; Young, supra
note 31, at 342-46.
84 See FED. R. EVID. 801(d)(2).
S2
BROOKLYN LAW REVIEW
[Vol. 60: 725
provides a defendant with little incentive to proceed to trial
once an offense has been accounted for in the original sentencing hearing. An acquittal saves little beyond a more extensive
record. For many defendants not wealthy enough to afford
multiple trials and not poor enough to qualify for a federal
defender, avoiding such collateral consequences may not be
worth the expense.
In addition to protecting the individual, the Double Jeopardy Clause safeguards important systemic finality goals.85
The Clause preserves the sanctity of judicial decisions, thus
conserving precious judicial resources and ensuring that "court
proceedings... 'command the respect and confidence of the
public.' 8' 6 "By preventing harassment and inconsistent results," the Double87 Jeopardy Clause protects the criminal justice system itself.
Prior-conviction enhancements are entirely consistent with
the finality of the original verdict. The verdict is in no way disturbed; facts found by judge or jury remain intact. It is the fact
of the first conviction, as opposed to the underlying events
leading to it, that merits additional punishment. System-wide
efficiency and consistency goals are not implicated.
In the case of the unadjudicated-offense enhancement, the
finality value is wholly offended. The punishment-enhancement premise exposes the system to successive revisions of
findings of fact. The significance of events comprising relevant
conduct may be repeatedly relitigated. The Guidelines require
all relevant conduct to be accounted for at sentencing regardless of whether the unadjudicated conduct has been the subject
of enhancement at another proceeding.8 Thus, one district
court may reject entirely the government's accusations, while
another may accept the same information with little investigation. Under the Guidelines, such revisions are not restricted to
the relatively ad hoc judgments at sentencing but extend to
judgments of acquittal after a full and fair jury trial.8 9 Not
See FRIEDLAND, supra note 77; Thomas, supra note 3, at 869.
FRIEDLAND, supra note 77, at 4 (quoting Connelly v. D.P.P., 1964 App. Cas.
1254, 1353 (appeal taken from Criminal Ct. of Appeals)).
"
87 FRIEDLAND, supra note 77, at 4.
See U.S.S.G. § 5G1.3 comment.
The now-famous colloquy between the district judge and defense counsel in
United States v. Juarez-Ortega, 866 F.2d 747, 748-49 (5th Cir. 1989), provides the
"
"
1994]
DOUBLE JEOPARDY
only does this erode the credibility of the judiciary, but it further subverts the efficiency of a badly overtaxed system."
It is time to abandon the fictional view of punishment
most fascinating (and horrifying) illustration of the dangers of acquittal enhancements. In Juarez-Ortega, the defendant was acquitted of use of a firearm in connection with a drug offense, but the district court justified an upward departure
on the basis of the gun use. For the benefit of those unfamiliar with the passage,
it is reprinted below:
The Court:
The jury could not have made-the jury could not have listened to the instructions.
[Counsel]:
Your Honor,The Court:
The testimony was so strong. The gun was even in the
apartment. That's all they needed. There was no dispute of
that fact. The mere fact that that gun was in the apartment, being used in association with-he didn't have to
have it on his person.
[Counsel]:
They perhaps didn't believe it was being used in association with drug-related activity, your Honor.
The Court:
Well, I'll tell you something: I have been disappointed in
jury verdicts before, but that's one of the most important
ones, because what it did, it set up a disparity in result
between the two defendants. Your client was consistently
selling cocaine from his apartment and using a firearm.
The fact is that the officers came in and testified that it
was in your clients waistband and described, had an officer
on the stand, a man who is an ATF agent, who is capable
and knows what a firearm looks like, telling them, "This is
what I saw."
There is no reason for him not to have seen that, since its
undisputed that the firearm was in the apartment and its
undisputed that the firearm was used in connection with
drug sales and used [for] the purpose of protecting drug
sales. And then here in number twelve, there is no doubt
at all that the firearm was brought for him. It's all a pattern. This firearm was used. They had to absolutely disregard the testimony of a government agent for no reason-no reason.
[Counsel]:
Perhaps they considered the testimony of the other agent
who testified that he couldn't be sure, your Honor.
The Court:
Well, you can take it up with an appellate court because
I've made my findings on the record.
Id. at 749. On appeal, the Fifth Circuit found nothing objectionable about the
judge's decision. Id.
' During the years between 1980 and 1990, criminal prosecutions in the federal courts for drug offenses increased by 258%. U.S. DEP'T OF JUSTICE,
SOURCEBOOK OF CRIMIINAL JUSTICE STATISTICS: 1992, at 485 (Kathleen McGuire et
al. eds.). In 1991, each district court had an average of 71 pending cases. Id. at
485. In addition, a post-Guidelines survey revealed that 50% of federal judges
responding estimated a 25% increase in time devoted to sentencing and one-third
of those surveyed reported an increase of 50%. See FEDERAL COURTS STUDY COMMISSION, REPORT OF THE FEDERAL COURTS STUDY CONMISSION 137 (1990).
BROOKLYN LAW REVIEW
[Vol. 60: 725
adopted to explain the use of prior convictions at sentencing.
The glare of the Guidelines exposes the flaws in the approach.
Both prior-conviction enhancements and enhancements for
unadjudicated crimes punish the defendant by causing her sentence to be longer than it otherwise would be. But the rejection
of unadjudicated offenses at sentencing does not require us to
ignore criminal history. Additional punishment upon a second
conviction is consistent with the language, history and underlying values of the Double Jeopardy Clause. Significantly,
enhancements for unadjudicated crimes have no such heritage,
are explicitly prohibited by affirmative requirements in other
clauses of the Constitution,9' and are wholly inconsistent with
the most basic goals of the double jeopardy provision.
McCormick was correct in shunning precedent and identifying
related offense enhancements as punishment.
III. DOUBLE JEOPARDY AND THE SUBSEQUENT-PROSECUTION
DILEMMA
Jettisoning the punishment-enhancement distinction does
not magically transform the Double Jeopardy Clause into a
powerful weapon at sentencing. In fact, a close analysis of the
Second Circuit's attempt to apply prevailing double jeopardy
doctrine to the subsequent-prosecution scenario demonstrates
the extent to which that doctrine must be perverted to achieve
even a moderate degree of justice. The major difficulty stems
from the fact that double jeopardy jurisprudence is predicated
upon the existence of a triggering event: a second trial after
conviction, acquittal or mistrial; or a second punishment after
an earlier conviction. Real-offense sentencing evades the trial
and conviction requirements, thereby positioning punishment
at sentencing outside the literal scope of the Double Jeopardy
Clause.
Double jeopardy challenges typically are characterized as
either "multiple punishment" or "successive prosecution"
claims92 because a defendant may neither be punished nor
See infra text accompanying notes 130-35; Lear, supra note 1, at 1223-37.
See United States v. Dixon, 113 S. Ct. 2849, 2881 (1993) (Souter, J., concurring and dissenting in part) (explaining that the Double Jeopardy Clause protects
against "multiple punishment" and "successive prosecution").
',
9
DOUBLE JEOPARDY
19941
prosecuted more than once for the same offense. Although the
Court has flirted with a broader "offence" definition under the
Double Jeopardy Clause,93 the Blockburger test supplies the
classic method for determining whether two offenses are in fact
the same.94 Under the Blockburger formula, a subsequent
prosecution is constitutionally acceptable so long as each statutory offense
"requires proof of a fact which the other does
95
not.)
Moreover, in the case of a multiple punishment claim,
even "the Blockburger test will not prevent multiple punishment where legislative intent to the contrary is clear."96 Ac-
cording to the Supreme Court in Missouri v. Hunter, "[w]here
... a legislature specifically authorizes cumulative punish-
ments under two statutes, regardless of whether those two
statues proscribe the same conduct under Blockburger, a
court's task of statutory construction is at an end and the [government may obtain] cumulative punishment under such statutes in a single trial."97 Thus, although the Blockburger test
will prevent the government from imposing a second punishment for the same offense in a subsequent prosecution, the
government may in a single proceeding impose multiple punishment for the "same offense" so long as legislative intent to
do so is absolutely clear. Absent a plain expression of intent,
however, the Double Jeopardy Clause prohibits the imposition
of additional punishment for that offense regardless of whether
the new sentence runs concurrently with the one originally imposed. 8 In Ball v. United States, the Court rejected concurrent sentences as a remedy for multiple punishment violations
on the ground that an additional conviction carries with it the
" See Grady v. Corbin, 495 U.S. 508 (1990) (adopting conduct-based offense
definition); Dixon, 113 S. Ct. at 2860 (overruling Grady and adopting the
Blockburger test to determine if offenses are the "same").
" Blockburger v. United States, 284 U.S. 299, 304 (1932).
g Dixon, 113 S. Ct. at 2866 (Rehnquist, C.J., concurring and dissenting in
part) (quoting Blockburger, 284 U.S. at 304).
96 Id.
at 2882.
459 U.S. 359, 368-69 (1983).
9,See Whalen v. United States, 445 U.S. 684, 694-95 (1980); Peter Westen,
The Three Faces of Double Jeopardy: Reflections on Government Appeals of Criminal Sentences, 78 MICH. L. REV. 1001, 1031-32 (1980); Kenneth G. Schuler, Note,
Continuing Criminal Enterprise Conspiracy and the Multiple Punishment Doctrine,
91 MICH. L. REV. 2220, 2226-27, 2232 (1993).
BROOKLYN LAW REVIEW
[Vol. 60: 725
99
possibility of "adverse collateral consequences."
A. McCormick's Multiple Punishment Theory
Although the second and third prongs of the McCormick
approach seem significantly less controversial than the first,
they illustrate the serious difficulties with invoking the double
jeopardy prohibition in the sentencing setting.'0 0 Following a
classic multiple punishment analysis, the second step of the
McCormick test asks whether Congress intended for a defendant to be punished once through relevant conduct enhancements, and again after conviction in a subsequent prosecution.
The question itself is flawed, however, because the court assumes that Congress may authorize cumulative punishments
for offenses deemed the same under the Blockburger test to be
administered in successive proceedings.'0 ' To date, the Supreme Court has limited cumulative punishment for "overlapping" offenses to a single trial setting.' 2 Even if the legislature permits multiple punishment for the same offense, the
multiple prosecution component of the Double Jeopardy Clause
bars multiple trials for offenses failing the Blockburger test.
But because courts have refused to equate the sentencing hearing with a prosecution,0 3 the multiple prosecution protection
does not provide the usual limit on the legislature's power in
this context. Thus, under the McCormick approach, the legislature is free to prescribe cumulative punishment for the same
offense through a series of sentencing hearings or trials.
Also troubling is the fact that the Second Circuit reviewed
the Guidelines-themselves for evidence of congressional intent
rather than the Sentencing Reform Act. Although cumulative
punishment may be authorized by the legislature, such autho-
" 470 U.S. 856, 865 (1985). These collateral consequences include "social stigma," "parole ineligibility," "impeachment of character in a future proceeding," and
"possible increased sentence pursuant to a recidivist statute." Id.
100See supra text accompanying notes 45-53.
101 United States v. McCormick, 992 F.2d 437, 439 (2d. Cir. 1993).
102 See Missouri v. Hunter, 459 U.S. 359, 369 (1983).
103 See, e.g., United States v. Koonce, 885 F.2d 720, 722 (10th Cir. 1989) (Al-
though defendant had arguably been punished for the unadjudicated offense at
sentencing, he had not been charged with the offense at issue and the multiple
prosecution prohibition was therefore not implicated.), cert. denied, 112 S. Ct. 1695
and 112 S. Ct. 1705 (1992).
1994]
DOUBLE JEOPARDY
rization must be unequivocally clear. 0 4 Absent evidence of
such intent, the rule of lenity requires courts "to resolve doubts
in the enforcement of a penal code against the imposition of
harsher punishment."0 5 Realistically, the Guidelines represent the intent of the Sentencing Commission rather than of
Congress. Congress takes no affirmative steps regarding the
content of the Federal Sentencing Guidelines. The Guidelines
themselves are drafted by the Commission and become law
automatically, absent congressional disapproval.' 6
A review of the Sentencing Reform Act' 7 confirms that
the legislation sought to eliminate "unwarranted sentence
disparities among defendants with similar records... [convicted] of similar crimes. " 'O' The Act also stressed "the appropriateness of imposing an incremental penalty for each offense
in a case in which a defendant is convicted of... multiple offenses committed in the same course of conduct." 09 On the
other hand, 18 U.S.C. § 3584 authorizes the imposition of consecutive or concurrent sentences in cases where the defendant
"is already subject to an undischarged term of imprisonment,"
and seems to require consecutive sentences when "multiple
terms of imprisonment are imposed at different times" unless
otherwise ordered by the court."0 This general authorizing
language seems a bit vague to overcome the presumption in
favor of noncumulative punishment, but had the Commission
embraced a consecutive-sentence approach, it would at least be
arguable that Congress had authorized such a result. Under
the Second Circuit's approach, then, a Commission decision to
authorize cumulative punishment for offenses punished
through relevant conduct and later prosecuted in a separate
proceeding, would be respected even though current multiple
punishment doctrine forbids cumulative punishment in succes-
,o See Peter Westen & Richard Drubel, Toward a General Theory of Double
Jeopardy, 1978 SuP. CT. REV. 81, 119-20; Schuler, supra note 98, at 2232.
05 Bell v. United States, 349 U.S. 81, 83-84 (1955); Westen & Drubell, supra
note 104, at 117-18.
6 28 U.S.C. § 994 (1988).
"0 Sentencing Reform Act of 1984, 18 U.S.C. §§ 3551-3586 (1988), 28 U.S.C. §§
991-998 (1988).
108 18 U.S.C § 3553(a)(6) (1988); S. REP. No. 225, 98th Cong., 1st Sess. 38
(1984).
1- 28 U.S.C. § 994(l)(1)(A).
110 18 U.S.C. § 3584(a).
BROOKLYN LAW REVIEW
[Vol. 60: 725
sive proceedings.
More to the point, however, I do not read the Guidelines to
authorize cumulative punishment in the McCormick situation.
As the dissent in McCormick pointed out, section 5G1.3, entitled "Imposition of a Sentence on a Defendant Subject to an
Undischarged Term of Imprisonment," theoretically avoids disparity in sentences brought about by piecemeal prosecution.
The relevant subsection for our purposes is section 5G1.3(b),
which provides
If subsection (a) does not apply, and the undischarged term of imprisonment resulted from offense(s) that have been fully taken into
account in the determination of the offense level for the instant
offense, the sentence for the instant offense shall be imposed to run
concurrently to the undischarged term of imprisonment.'
According to the background commentary, the "guideline is
intended to result in an appropriate incremental punishment
for the instant offenses that most nearly approximates the
sentence that would have been imposed had all the sentences
been imposed at the same time.""2 The guideline and accompanying commentary are hardly models of clarity. 1 '3 The "fully taken into account" language is somewhat obscure; however,
the application notes suggest that it is limited to situations
involving related offenses factored in as part of relevant conduct."' Related offenses accounted for as specific offense
"I U.S.S.G. § 5G1.3(b). Note that subsection (b) applies only if subsection (a)
does not. Subsection 5G1.3(a) applies to offenses committed while defendant was
"serving a term of imprisonment . .. or after sentencing for, but before commencing service of, such term of imprisonment." In such a case, the Guideline requires
that the sentence for the instant offense be "imposed to run consecutively to the
undischarged term of imprisonment." U.S.S.G. § 5G1.3(a). Subsection 5G1.3(c) is a
policy statement that allows consecutive sentences in cases not covered by (a) or
(b), to the extent "necessary to achieve a reasonable incremental punishment for
the instant offense."
" See U.S.S.G. § 5G1.3, comment. (backg'd).
11 The guideline has been amended four times since it was first promulgated in
1987. On several of these occasions it was deleted its entirety and replaced with a
new guideline dealing with the same issue.
.. Application Note 2 provides in pertinent part: "Subsection (b) .
.
. addresses
cases in which the conduct resulting in the undischarged term of imprisonment
has been fully taken into account under § 1B1.3 ([rlelevant conduct) in determining the offense level for the instant offense. U.S.S.G. § 5G1.3, comment. (n.2)." It
is not at all clear that subsection (b) requires concurrent sentences when related
offense have been considered via relevant conduct, but the full force of the offense
has been limited by the statutory maximum for the offense of conviction. In such
1994]
DOUBLE JEOPARDY
characteristics, 5 or through another adjustment to the offense level, do not appear to be included. Such enhancements
are not generally equivalent in length to the sentence that
would have followed a conviction for that offense. These offenses would therefore not be "fully take[n] into account" in an
original sentence.
A fair reading of section 5G1.3 in the McCormick case
should have produced a sentence that, when combined with the
earlier sentence, would have roughly approximated the total
sentence the defendant would have received had he been tried
on all counts in one proceeding. The application notes appear
to require the district court to (1) calculate the guidelines sentence using all relevant conduct, including conduct for which a
conviction has been obtained; (2) calculate the sentence for the
instant offense; (3) subtract from this number the amount of
time already served for the related offense; and (4) enter a
sentence that runs concurrently
with the remaining time left
6
on the first sentence."
Thus, assuming that McCormick had been convicted on all
forty-one counts of the Vermont indictment, the court should
have used the Connecticut frauds to determine the base offense
level, added in the appropriate adjustments, and chosen a
sentence. Assuming for our purposes that McCormick had
served six months of his forty-six-month sentence, the Vermont
court should have subtracted six months from the total Versituations, subsection (c) appears to allow an appropriate consecutive sentence.
Relevant conduct is not the only method by which related offenses are accounted for at sentencing. Section 5G1.3 does not appear to demand concurrent
sentences for offenses taken into account as specific offense characteristics, as the
basis of an obstruction enhancement, see U.S.S.G. § 3C1.1, or through other offense-level adjustments. Adjustments of this sort normally do not yield offense
level increases equal in length to the sentence that would have followed a formal
conviction for that offense. Thus, unadjudicated crimes included as specific offense
characteristics or through other adjustment provisions do not satisfy the "fully
take[n] into account" requirement in § 5G1.3(b).
...For example, under U.S.S.G. § 2D1.1(b)(1) the base offense level of a defendant convicted of distributing a controlled substance should reflect a two-level
increase if a dangerous weapon was possessed as part of the drug transaction. Use
of a gun during a drug transaction constitutes a separate criminal offense under
18 U.S.C. § 924(c). Because the application of § 5G1.3(b) appears to be restricted
to relevant conduct enhancements, a sentence based upon a subsequent conviction
for the gun offense would run consecutively and would not credit the prison time
attributable to the gun enhancement.
116See U.S.S.G. § 5G1.3(b) comment. (n.2).
BROOKLYN LAW REVIEW
[Vol. 60: 725
mont sentence and entered a concurrent sentence as to the
rest. The result should have equaled the aggregate prison term
McCormick would have received had the Connecticut and Vermont counts been joined in a single indictment.
The McCormick panel, however, concluded that the subsequent prosecution, if successful, would impose an additional
punishment unauthorized by Congress, and went on to consider whether concurrent sentencing (like that potentially
authorized by section 5G1.3) would avoid the double jeopardy
bar. In Ball v. United States, the Supreme Court acknowledged
that a defendant convicted of multiple, overlapping offense
could suffer "adverse collateral consequences."' 17 Such consequences included the possibility that the additional conviction
would be used to enhance his sentence under a recidivist statute, impeach his testimony at a later proceeding, or affect a
parole decision." 8
Here again the traditional multiple punishment analysis
breaks down in the sentencing setting. All of the adverse collateral consequences contemplated in Ball derive from the
existence of the additional conviction for the same offense, as
opposed to any additional punishment. In the McCormick scenario, however, there is no second conviction from which the
collateral consequences might flow. Although the same conduct
could conceivably be used against the defendant at any number
of sentencing hearings, only one conviction (as opposed to punishment) can constitutionally be imposed for that offense. In
the recidivist statute example, a defendant who commits a
subsequent state offense might suffer recidivist treatment if
that state counts only the number of convictions to determine
recidivist status." 9 But the federal government could constitutionally have prosecuted the offenses at separate proceedings
in the first instance. Unless we are prepared to say that such
collateral consequences present a double jeopardy bar to separate prosecutions of related offenses, it seems odd to suggest
that concurrent sentences in McCormick give rise to such a
17
See Ball v. United States, 470 U.S. 856, 865 (1985). For a discussion of
"collateral consequences," see Schuler, supra note 98, at 2227.
118 Ball, 470 U.S. at 865; Schuler, supra note 98, at 2227.
119 See Schuler, supra note 98, at 2228 n.59 (listing state statutes). In the federal system, the Guidelines count concurrent sentences arising from related offenses only once for criminal history purposes. U.S.S.G. § 4A1.2(a)(2).
1994]
DOUBLE JEOPARDY
bar.
B. Double Jeopardy Blues
Singling out the multiple punishment prong of the Double
Jeopardy Clause for application at sentencing raises other serious doctrinal difficulties. If the punishment finding is limited
to subsequent-prosecution scenarios like that in McCormick, 120 we find ourselves in an extremely awkward position
regarding acquittals: the McCormick solution attributes greater preclusive effect to the use of an unadjudicated offense at
sentencing than it does to an acquittal returned by a jury after
a full and fair trial. Thus, under the McCormick approach, the
multiple punishment prohibition would preclude the prosecution of an offense used at sentencing that has never been the
subject of an indictment and trial, but would create no bar to
punishment for acquittal conduct.
Attempting to harmonize the acquittal jurisprudence under the double jeopardy rubric opens a Pandora's box of problems. Traditional double jeopardy jurisprudence prohibits a
second trial after acquittal. Acquittals are thus protected under
the multiple prosecution prong as opposed to the multiple
punishment prong of the Double Jeopardy Clause. Certainly
the McCormick scenario fits easily into the multiple prosecution analysis,12" ' directly implicating the values protected un120 Even
adopting the limited exception for successive punishment cases crafted
by the Second and Tenth Circuits requires difficult interpretations of the use of
conduct at previous sentencing hearings. Consider, for example, United States v.
Garcia, 919 F.2d 881 (3d Cir. 1990). Simply figuring out whether-and in what
manner-the sentencing court in the previous prosecution had used the related offense information required four pages of detailed consideration. Id. at 883-87.
Peter Henning suggests that a better approach to the subsequent-prosecution
dilemma than that offered in McCormick or Koonce might be to focus on Dixon's
lesser-included offense analysis to determine whether the government is proving
the same elements that were introduced at sentencing to convict the defendant in
the subsequent prosecution. Peter Henning, Precedents in a Vacuum: The Supreme
Court Continues to Tinker with Double Jeopardy, 31 AM. CRIM. L. REV. 1, 30-31
(1993). This requires courts to distinguish between conduct that relates to the
quality of the first offense, and conduct that is outside the offense of conviction
but is considered to be related to the sentenced criminal activity. Id. at 31. If the
conduct concerns the quality of the offense, then its use in sentencing does not
raise double jeopardy concerns because it will not involve proof of elements that
constitute a separate offense. Id.
121 In fact, if one concludes, as I do, that U.S.S.G. § 5G1.3(b) avoids any addi-
BROOKLYN LAW REVIEW
[Vol. 60: 725
der the multiple prosecution prohibition. Though not inspiring
the ire of the acquittal cases, prosecutions for offenses already
punished as relevant conduct insult the central purposes of the
multiple prosecution prohibition. As pointed out earlier, one of
the main objects of the Double Jeopardy Clause is to avoid
repeated dress rehearsals and the attendant danger that a
defendant, though innocent, may be found guilty.12 2 Guidelines sentencing allows the government to punish a criminal
offense without the nuisance of indictment or trial, while al23
lowing the prosecution to perfect its case at sentencing.'
Not only does the Guidelines system leave open the real possibility that an innocent person may be punished through relevant conduct, but it enhances the prospect that an innocent
person may be convicted at a formal trial after the government
has had a "dry run" at sentencing. Such a system strikes at
the very heart of the multiple prosecution protection.'24
But applying the multiple prosecution prong to acquittal
enhancements or to the McCormick subsequent-prosecution
scenario requires courts to equate the sentencing hearing with
a trial or prosecution.'2 5 For obvious reasons, courts have
tional punishment after the subsequent prosecution, the multiple-prosecution prong
is the only means by which the Double Jeopardy Clause conceivably could apply.
1
See Green v. United States, 355 U.S. 184, 187-88 (1957).
12 In cases involving relevant conduct enhancements, this possibility is foreclosed in the Second and Tenth Circuits.
124 The rationale for extending double jeopardy protection to government-provoked mistrials applies with even greater force in the sentencing arena. Jeopardy
is held to attach at the beginning of trial, rather than upon conclusion, to prevent
the government from avoiding the double jeopardy bar by the simple expedient of
provoking a mistrial. Professor Westen and co-author Drubell contend that the
mistrial cases are explained by a combination of finality interests and the desire
to thwart government manipulation. See Westen & Drubell, supra note 104, at 9197. Under the Guidelines, the government may evade the reach of the Double
Jeopardy Clause simply by declining to indict the defendant for the offense at
issue. The government's ability to avoid the affirmative obligations of the Fifth
and Sixth Amendments likely exposes the system to substantially greater abuse
than does manipulation of the mistrial process. The likelihood of an inaccurate
conviction after mistrial is negligible compared to the possibility of an erroneous
judgment and punishment at sentencing.
1" The multiple prosecution prong protects a defendant from a second trial for
the same offense after conviction or acquittal. It does not protect the defendant
from a second proceeding, nor does it technically protect a defendant from punishment for the acquittal conduct after the acquittal.
the guaranLong established constitutional doctrine makes clear that ...
tee against double jeopardy imposes no restrictions upon the length of a
1994]
DOUBLE JEOPARDY
balked at taking such a step.12 It would be implausible to
suppose that the sentencing hearing could obtain such status
for double jeopardy purposes without provoking the application
of the entire2 7spectrum of procedural requirements, including
trial by jury.
Holding that enhancement equals punishment in the multiple punishment context is equally precarious. No principled
method for limiting the "punishment" designation to double
jeopardy cases exists. This finding is in direct conflict with
precedent in almost every circuit upholding the use of acquittals 128 and other uncharged conduct. 129 The question of how
the government constitutionally may "punish" anyone for a
criminal offense without complying with the Fifth and Sixth
Amendments seems certain to follow.
sentence imposed upon reconviction. At least, since 1896, it has been
settled that this constitutional guarantee imposes no limitations whatever
upon the power to retry a defendant who has succeeded in getting his
first conviction set aside. . . . And, at least since 1919, . . . it has been
settled that a corollary of the power to retry a defendant is the power,
upon the defendant's reconviction, to impose whatever sentence may be
legally authorized, whether or not it is greater than the sentence imposed after the first conviction.
North Carolina v. Pearce, 395 U.S. 711, 719-20 (1969) (citations omitted). But see
William J. Kirchner, Punishment Despite Acquittal: An Unconstitutional Aspect of
the Federal Sentencing Guidelines?, 34 ARIz. L. REV. 799, 819 (1992) ("facts relating to a prior acquittal should be seen as a special category that is protected by
the Double Jeopardy Clause"). The reason double jeopardy jurisprudence has not
encompassed these situations is that the Constitution does not recognize the possibility of their existence. The affirmative commands of the Fifth and Sixth Amendments are not optional; the government acts lawlessly by imposing punishment for
particular conduct without first obtaining a conviction for that offense. See Lear,
supra note 1, at 1232.
" See, e.g., United States v. Carey, 943 F.2d 44, 46 (11th Cir. 1991); United
States v. Mack, 938 F.2d 678, 681 (6th Cir. 1991); United States v. Koonce, 885
F.2d 720, 722 (10th Cir. 1989).
127 See, e.g., United States v. Mobley, 956 F.2d 450, 455 (3d Cir. 1992) (refusing
to give a convicted criminal the "full panoply of constitutional rights" afforded to
an accused at the trial stage).
128 See, e.g., United States v. Rodriguez-Gonzalez,
899 F.2d 177, 182 (2d Cir.
1990); United States v. Mocciala, 891 F.2d 13, 17 (1st Cir. 1989); United States v.
Juarez-Ortega, 866 F.2d 747, 749 (5th Cir. 1989).
1'
See, e.g., United States v. Galloway, 976 F.2d 414, 422-27 (8th Cir. 1992),
cert. denied, 113 S. Ct. 1420 (1993); United States v. Martinez, 924 F.2d 209, 211
(11th Cir. 1991).
BROOXLYN LAWREVIEW
[Vol. 60: 725
IV. REJECTING THE REAL-OFFENSE MODEL
The problem with the McCormick approach to the subsequent-prosecution dilemma is that the court chose the wrong
remedy to deal with a legitimate constitutional problem. As the
foregoing analysis demonstrates, the Double Jeopardy Clause
is strangely ill-equipped to deal with the prosecution of offenses already used at sentencing or with the reconsideration of acquittal conduct for enhancement purposes. Instead of contorting double jeopardy jurisprudence to prohibit subsequent-prosecutions, courts should look to the Due Process Clause to prohibit the punishment of unadjudicated crimes in the first instance.
I have argued elsewhere that the systemic imbalance created by the real-offense premise renders the system unconstitutional under the Due Process Clause. 3 ' Consider the end
result of real-offense sentencing: an individual is punished for
an offense of which he has never been convicted. Allowing the
government to impose punishment for a criminal offense absent conviction dislodges the jury from its crucial oversight role
in the criminal justice system. In our constitutionally created
criminal justice regime, any of the three actors-judge, jury
(grand or petit) or prosecutor-acting alone may exercise mercy.'3 ' The reverse is not true. The Constitution requires una-
...See Lear, supra note 1, at 1186. The premise of the earlier article is that
real-offense sentencing is unconstitutional, not because of a change in sentencing
approach, but because it was unconstitutional in the first instance. Thus, I contend
that Williams v. New York, 337 U.S. 241 (1949), the opinion in which the Supreme Court gave its official blessing to real-offense sentencing, was wrongly decided. In addition, though I doubt the Court was correct 45 years later in
McMillan v. Pennsylvania, 477 U.S. 79 (1986), which held that the Due Process
Clause allows serious sentencing enhancements upon proof by a preponderance of
the evidence, I argue that it is the use of the nonconviction offense, rather than
the quantum of proof, that renders the McMillan scenario unconstitutional.
In questioning the legitimacy of the real-offense system, I join many other
scholars writing both before and after the adoption of the Guidelines. See, e.g.,
Heaney, supra note 31, at 185-225; Sanford H. Kadish, Legal Norm and Discretion
in the Police and Sentencing Processes, 75 HARV. L. REV. 904, 920-24 (1962), reprinted in SANFORD H. KADISH, BLAME AND PUNISHMENT 250 (1987); Reitz, supra
note 1; Stephen J. Schulhofer, Due Process at Sentencing, 128 U. PA. L. REV. 733,
757-72 (1980); Michael H. Tonry, Real Offense Sentencing: The Model Sentencing
and Corrections Act, 72 J. CRIM. L. & CRIMINOLOGY 1550, 1564-85 (1981).
3' The judge may dismiss the indictment or enter a judgment of acquittal absent the agreement of prosecutor or jury; the prosecutor may refuse to indict in
19941
DOUBLE JEOPARDY
nimity; all three actors must agree before an individual may be
punished for a criminal offense." 2 At the heart of this system
lie the grand and petit juries, haled by the people as the bulwark of liberty, the soul of American freedom.'33 Yet the realoffense system allows the government to obtain punishment
without ever presenting the charge for grand jury review, and
without ever preparing a case for petit jury scrutiny.1 4 The
three-way dialogue so critical to the stability of the criminal
justice system becomes linear; punishment decisions are the
result of discussions between judge and prosecutor. The affirmative commands of the Fifth and Sixth Amendments become
superfluous, and the jury-the crown jewel of our constitutional plan for the administration of justice-is rendered powerless
to protect the citizen from the overzealous prosecutor or the
unscrupulous judge.3 5
The subsequent-prosecution dilemma under the Guidelines
simply provides another example of the systemic dislocation
created by the real-offense system. Allowing punishment absent conviction renders the Double Jeopardy Clause incompetent to protect its most basic purposes.3 6 Instead of attempting to revamp various constitutional provisions to inject a
spite of opposition from the grand jury or judge; the grand jury may decline to
indict, and the petit jury decline to convict, despite the strength of the
government's case or the opinion of judge or prosecutor. See Lear, supra note 1, at
1223-24; see also Akhil R. Amar, The Bill of Rights as Constitution, 100 YALE L.J.
1131, 1184-85 (1991).
"2 Amar, supra note 131, at 1184-85; Lear, supra note 1, at 1224.
11 See FRANCIS H. HELLER, THE SIXTH AMENDMENT TO THE CONSTITUTION OF
THE UNITED STATES 21-29 (1951); SHANNON C. STISON, THE AMERICAN REVOLUTION IN THE LAW 34 (1990); 3 JOSEPH STORY, COMIENTARIES ON THE CONSTITUTION OF THE UNITED STATES § 1779, at 652-53 (Leonard W. Levy ed., Da Capo
Press 1970) (1833).
13 Many commentators have pointed out that the Guidelines' real-offense system actually makes it much easier for the prosecution to secure punishment for a
particular offense at sentencing than to obtain a conviction via trial or guilty plea.
For instance, a prosecutor faced with proof problems, either regarding the credibility of an informant or the presence of illegally seized evidence, may decline to
seek an indictment or drop the count before trial or plea. If that offense falls
within the scope of relevant conduct, she may in many instances obtain the identical sentence for that offense that she would have obtained had the defendant been
convicted. See Freed, supra note 26, at 1714; Gerald W. Heaney, Revisiting Disparity: Debating Guidelines Sentencing, 29 AM. CRIM. L. REV. 771, 774 (1992).
...See Duncan v. Louisiana, 391 U.S. 145, 156 (1968).
...See, e.g., United States v. Juarez-Ortega, 866 F.2d 747 (5th Cir. 1989) (judge
disagreed with jury acquittal and imposed punishment accordingly).
BROOKLYN LAW REVIEW
[Vol. 60: 725
modicum of fairness into sentencing, the courts should reconsider the real-offense regime in its entirety.
In the face of a practice that is demonstrably incompatible
with the fundamental principles of our criminal justice system,
the federal judiciary steadfastly claims the "discretion" to "enhance" for unadjudicated crimes. The real-offense component of
the Guidelines allows prosecutors to avoid the rigors of indictment and trial, is patently unfair, and generates an enormous
workload for the courts. 13 7 The federal courts lost control over
sentencing originally because they allowed an inequitable system to proceed unchecked. The courts should think hard about
the consequences of doing that twice.
The Commission, too, should reconsider its infatuation
with the real-offense model. Conduct forming the underlying
basis for a separate statutory offense, or conduct that, when
combined with the conviction-offense conduct, would constitute
a separate crime, 3 ' should be excluded from the sentencing
inquiry. 39 I recognize that this formula is more easily stated
than applied. Congress has not neatly defined federal crimes;
1-
See FEDERAL COURTS STUDY COMMISSION, supra note 90, at 137 (reporting
that over half of federal judges responding to the survey estimated a 25% increase
in time devoted to sentencing and one-third of those surveyed reported an increase
of 50%).
,"' Professors Tonry and Coffee refer to such facts as "legally recognized facts,"
which they define to be facts that form "elements of substantive offense definitions
and constitute the basis for hierarchical distinctions between offenses." Michael
Tonry & John C. Coffee, Jr., Enforcing Sentencing Guidelines: Plea Bargainingand
Review Mechanisms, in ANDREW VON HIRSCH ET AL., THE SENTENCING COMMISSION
AND ITS GUIDELINES 142, 155 (1987). For example, the Guidelines treat the use of
a gun in connection with a drug offense as a specific offense characteristic under §
2D1.1(b)(1), even though such conduct is a criminal offense under 18 U.S.C. §
924(c). If a conviction-offense regime were adopted that excluded both
nonconviction offenses and "legally cognizable facts," the gun could not be used as
a sentencing factor to calculate the appropriate sentence for a drug conviction.
Similarly the iuse of a gun could not be introduced as an aggravating factor
against a defendant convicted of robbery in a jurisdiction that has an armed robbery statute. See id. at 155-56.
...Professor Reitz proposed the following "working" definition of a convictionoffense system:
"Conviction-offense sentencing" allows punishment based on:
(1) The current conviction offenses,
(2) Aggravating and mitigating facts associated with the conviction
offenses, excluding facts that have been defined by the legislature as separate offenses or as elements of separate offenses, and
(3) Prior conviction offenses.
Reitz, supra note 1, at 535.
1994]
DOUBLE JEOPARDY
rather, overlapping statutes are the norm, and perhaps we
must accept that any rational sentencing system will require
the consideration of some extraneous, and arguably criminal,
conduct. Delineating the instances in which sentencing factors
should encompass criminal conduct outside the offense of conviction will require a more detailed analysis than is possible
here. Several commentators have suggested workable models
exhibiting greater respect for constitutional values and reflecting more
thoughtful policy choices than the Guidelines sys140
tem.
The Commission must recognize that the use of unadjudicated offenses at sentencing likely has contributed to, rather
than prevented, unwarranted disparity.' Too many factual
.. For example, Professor Reitz outlines a "beginning theory" of sentencing facts
and conviction facts in a conviction-offense system. According to Professor Reitz
"inquiries with the greatest claim on the sentencing process are those which: (1)
are necessary to effectuate a jurisdiction's purposes of punishment and (2) are too
detailed or qualitative to be captured in terms of a statute and decided at trial."
Reitz, supra note 1, at 566. Nonconviction offenses, legally cognizable facts (i.e.,
facts that form "elements of substantive offense definitions and constitute the basis
for hierarchical distinctions between offenses," Tonry & Coffee, Jr., supra note 138,
at 155), and "noncriminal 'biographic' conduct, separable from the conviction offense" are excluded from the sentencing inquiry. See Reitz, supra note 1, at 56571.
Professor Yellen has proposed a different approach to reform that would require "the Sentencing Commission to reconsider both the weight given to alleged
related-offenses and the categories of offenses that receive such treatment." Professor Yellen suggests that the Commission might "strike a different balance," by
discounting the penalty associated with a related offense at sentencing. The constitutional problem would remain, albeit in a less offensive form. Professor Yellen
also points out that decreasing the weight of related offenses in the sentence calculation might alleviate some of the concerns regarding proportionality and plea
bargaining which arise in a pure charge offense system, while making a plea bargain less illusory than it is under the current regime. Yellen, supra note 1, at
457-58.
141 The problems generated by the Guidelines system are a much greater threat
to uniformity than the evils the Commission sought to avoid. As Professor Yellen
explains, "the complexity of the relevant conduct standard . . . contributes to unwarranted disparity. Complexity invites errors in application or inconsistent interpretation of key concepts." See Yellen, supra note 1, at 451.
The real-offense components of the Guidelines were included to blunt the
effect of prosecutorial charging and bargaining leniency. See Ilene H. Nagel &
Stephen J. Schulhofer, A Tale of Three Cities, 66 S. CAL. L. REv. 501, 505 (1992).
Relevant conduct and similar innovations combat this problem by assuring appropriate severity. But I remain unconvinced that prosecutorial leniency is a pervasive phenomenon or that charge bargains routinely underrepresent a defendant's
true culpability. Federal prosecutors are self-selected to prosecute. They are rewarded not for soft-hearted dispositions of serious cases, but for convictions. See
BROOKLYN LAW REVIEW
[Vol. 60: 725
judgments must be made by too many individual decisionmakers to produce consistent results.'
District judges use a
spectrum of proof requirements.' The personal experience of
the district judge undoubtedly influences her acceptance of
various types of testimony. One court may routinely reject drug
weight approximations based upon hearsay, while another may
repose complete confidence in such testimony. In contrast to
jury deliberation, eleven other viewpoints do not challenge or
refine the decision. And although findings of fact in the district
courts generate dramatic fluctuations in sentence lengths, the
Elizabeth T. Lear, Contemplating the Successive Prosecution Phenomenon in the
Federal System, 85 J. CRIM. L. & CRIMINOLOGY (forthcoming 1995). Many politicians began as prosecutors, and the voting public consistently rewards the "toughon-crime" stance. As such, the threat of the lenient prosecutor is overrated.
Nor does it seem likely that prosecutors routinely agree to plea bargains that
seriously underrepresent the defendant's culpability. Years of documented overcharging should undermine that conclusion. Why assume that prosecutors always
dismiss easily provable charges rather than offenses that they know will be difficult to prove? Defendants are regularly acquitted in the federal courts; not every
indictable offense can be proven at trial. There likely exists a group of cases in
which an Assistant U.S. Attorney dismisses or declines to indict an offense because she believes that it cannot be proven at trial. Yet, the Guidelines freely
"count" such offenses, thus overriding legitimate exercises of prosecutorial discretion.
142 See Freed, supra note 26, at 1715 (district judges, prosecutors, defenders and
probation officers demonstrate widely different attitudes and practices respecting
relevant conduct within and across districts); Yellen, supra note 1, at 403.
14 Although every circuit accepts the notion that most sentencing factors need
only be proven by a preponderance of the evidence to warrant enhancement, I
doubt seriously that "proof by a preponderance" is sufficiently specific to elicit consistent findings based upon the same evidence across the board. See, e.g., United
States v. Hendricks, 956 F.2d 1164 (6th Cir. 1992) (requiring "some evidentiary
basis beyond mere allegation in an indictment), cert. denied, 113 S. Ct. 139 (1992);
United States v. West, 948 F.2d 1042, 1045 (6th Cir. 1991) (requiring "evidence of
other quantities of drugs involved must have a minimal level of reliability beyond
mere allegation"), cert. denied, 112 S. Ct. 1209 (1992).
Consider the South Dakota District Court's decision regarding the appropriate
drug amount in United States v. Koonce. Koonce's customer testified that he had
purchased approximately twenty to thirty pounds of methamphetamine during the
preceding year. 945 F.2d 1145, 1148 n.2 (10th Cir. 1991), cert. denied, 112 S. Ct.
1695 and 112 S. Ct. 1705 (1992). Naturally suspicious of such an estimate from a
cooperating drug defendant, the court held Koonce accountable for only 14.22
pounds. See id. The court may well have sentenced Koonce to a much greater
term than he deserved because the judge's estimate was inflated, or to a substantially shorter term because the amount underrepresented the true sales. Such decisions are highly unreliable and it seems likely that similarly situated sellers will
receive drastically different terms of imprisonment pursuant to such an ad hoc
system.
DOUBLE JEOPARDY
19941
courts of appeals review such decisions for clear error,'" thus
eliminating any possibility of even intra-district consistency.
The entire enterprise demands the vigilance of the probation officer and the cooperation of the prosecutor. But probation officers do not evaluate relevant conduct information consistently.'4 5 Thus, the sentencing range recommended by the
probation department, and often adopted summarily by the
court, may vary dramatically depending upon the officer. And
probation officers often operate in the dark because prosecutors
do not always divulge pertinent information. In addition to
withholding critical facts from the probation officer, prosecutors retain a myriad of options through which to manipulate
the sentences under the Guidelines. As former Commissioner
Nagel and Professor Schulhofer report, prosecutors can and do
circumvent the Guidelines through plea bargaining. 4 ' Prosecutors may shield defendants from the full force of relevant
conduct by charging offenses that have low statutory maxima 47 or are outside the scope of the relevant conduct provision. 4" Note also that the definition of "substantial assistance" 49 varies by district 50 and invites further unwarranted disparity. Even if a U.S. Attorney creates a clear and consistent policy regarding substantial assistance motions, the
extent of the departure cannot be controlled. The real-offense
'" See, e.g., United States v. Koonce, 884 F.2d 349, 353 (8th Cir. 1989).
14 See Pamela B. Lawrence & Paul J. Hofer, An Empirical Study of the Application of Relevant Conduct Guideline § 1B1.3, 4 FED. SENT. REP. 330 (1992).
.. See Nagel & Schulhofer, supra note 141, at 557.
...See Yellen, supra note 1, at 440; Nagel & Schulhofer, supra note 141, at
547.
14
See Yellen, supra note 1, at 440.
,' U.S.S.G. § 5K1.1 authorizes the court to depart for "substantial assistance"
to authorities, but only upon a motion by the government.
.50Nagel & Schulhofer, supra note 141, at 550-51. According to the authors,
The use of... 5K1.1 substantial-assistance motion varies from jurisdiction to jurisdiction. The motion is subject to supervisory review, at
least in theory. There is no limit on the amount of reduction once the
motion is submitted. The section 5K1.1 motion is also used to avoid
guidelines ranges or mandatory minimum sentences for sympathetic defendants--even when there has been no genuine substantial assistance.
[Fjederal defenders ..
often claim that AUSAs refuse to file
the 5K1.1 substantial-assistance motion even after a defendant has fully
cooperated.
BROOKLYN LAW REVIEW
[Vol. 60: 725
regime simply creates too many opportunities for ad hoc judgments and manipulation to provide anything approaching
legitimate uniformity in sentencing."'
CONCLUSION
McCormick was correct in recognizing enhancement as
punishment, but failed to embrace the appropriate constitutional remedy. Instead of preventing the subsequent prosecution of the offense at issue, the court should have prohibited
the original sentence enhancement. The real-offense system
should be abandoned on constitutional and practical grounds.
It encourages the government to circumvent the affirmative
commands of the Fifth and Sixth Amendment. In addition, it
has spawned enormous administrative difficulties and has not
produced the expected benefits. Sentencing remains fraught
with inequities. Eliminating the real-offense character of the
Guidelines would avoid many of these inequities and bring us
closer to our goal of justice in sentencing.
151 The sentencing guidelines were also expected to inject some "honesty" into
sentencing. According to the Commission, this was easily achieved by abolishing
parole, thereby making the sentence imposed the sentence served. The public,
however, remains in the dark. Certainly, as the fact that parole no longer exists
filters through the general populace, citizens will begin to recognize that the sentence imposed is "real." Yet, consider that in McCormick the sentence resulting
from the subsequent prosecution would, for the most part, have been concurrent.
While the absence of § 5G1.3(b) would allow prosecutors to evade the uniformity
goal, the current approach allows the local U.S. Attorney to obtain "credit" in
cases in which the defendant has already been punished for the offense at issue.
Few members of the larger populace will understand that the sentence was imposed to run concurrently. And those who do will undoubtedly wonder why government resources were diverted for a redundant prosecution.
In addition, the public now sees defendants convicted of identical crimes who
routinely receive vastly different sentences. Though most people expect sentence
lengths to vary a bit, the remarkably dissimilar sentences produced by the Guidelines after identical convictions will eventually undermine the credibility of the
federal courts. Public confidence is particularly important among groups most likely to come into contact with the criminal justice system. Such groups are also the
most likely to suspect that wild fluctuations in sentences for similar crimes are
the result of bias and favoritism.